This is a good stopping point,

the two vcc2 pads aren't connected but it doesn't really matter. I'll ponder how to connect them. probably tweak it all some more but this is lookin good.

The solution for every routing problem is more layers!

it's not actually a problem, the last major version didn't even map those two, also I think the only cpu that may even have pins for it is k6 and I'm pretty sure it doesn't care. also the mobo links those pads anyway so it's all totally pointless worring about those two vcc2 pins.

in any case this is only useful for those trying to power vcc2 externally and at that it's not needed so. I decided leaving them floating (but connected through mobo) was best call.

okay. this seems to be a good prototype candidate. I scoured over this making small tweaks but don't see any real issues. may put in a order for this upload if I don't see anything else in coming days.

put in a order for a 0.6 fiber one, a 1 aluminum one, and a 1.6 aluminum one

still haven't gotten the first order 🤣

just got a email. looks like they are limited in what they can do on aluminum. this project is too complicated for aluminum. so gonna cancel the aluminum pcbs I ordered and use fiber

What would be the advantage for aluminum in this application?

Easier time shorting out your mobo and killing it comes to mind 🤣

This is pretty interesting, hope it works on first try ~

Kahenraz wrote on 2022-04-22, 11:13:

What would be the advantage for aluminum in this application?

probably nothing other than it's cool 🤣

EMI may be less with aluminum.

Tiido wrote on 2022-04-22, 12:13:

Easier time shorting out your mobo and killing it comes to mind 🤣

This is pretty interesting, hope it works on first try ~

the aluminum pcbs are supposedly rated to 3000v They seem legit. but the fab has more limitations in what they can make physically with them vs traditional pcbs. Apparently they have been used with LEDs for some time for heat dissipation.

It's very entertaining watching someone try to solder or deslolder from an aluminum PCB with a piddly little iron. They are basically giant heat sinks.

Kahenraz wrote on 2022-04-22, 14:07:

It's very entertaining watching someone try to solder or deslolder from an aluminum PCB with a piddly little iron. They are basically giant heat sinks.

Good point, I know I had trouble once with 2 oz solder and no thermal reliefs. that was hell 🤣.

Sphere478 wrote on 2022-04-22, 13:24:

the aluminum pcbs are supposedly rated to 3000v They seem legit.

Only on the top surface where tracks are made, but not in anywhere else. You won't get insulation on all of the outer surface, plus that insulation is easy to mechanically damage and that is what the shorting out comment was about. As soon as you lay it on the pins, there are gonna be penetrations at the edges of drilled holes and all hell breaks loose 🤣. Plus it is pain to solder, since heatsink is always bad to solder 🤣

Tiido wrote on 2022-04-22, 14:55:

Sphere478 wrote on 2022-04-22, 13:24:

the aluminum pcbs are supposedly rated to 3000v They seem legit.

Only on the top surface where tracks are made, but not in anywhere else. You won't get insulation on all of the outer surface, plus that insulation is easy to mechanically damage and that is what the shorting out comment was about. As soon as you lay it on the pins, there are gonna be penetrations at the edges of drilled holes and all hell breaks loose 🤣. Plus it is pain to solder, since heatsink is always bad to solder 🤣

Of course, it's Simply a insulation rating. "Other factors may apply." you're right.

They’re here!

-Pcb thickness at 0.6 looks nice and thin.

-the silk didn’t turn out well as I feared. But idk if there is much I can do, there is so much info to include.

Okay so now that I’m playing with these, it looks like there are going to be a few changes.

Something keeps happening where vss keeps getting shorted to vcc3 or vcc2
I’m not sure if this is a installation error or a wiring error, it is gonna take me a while with a volt meter to debug that. I was wondering if something like this may happen, I’m testing it with a volt meter on a non working mobo, and once the install looks good, I will run my processor collection through it and make sure that it doesn’t show any shorts there. If all looks well I’ll do another install on a fic 503+. A board which doesn’t require the help of this tweaker really except when it comes to the tillamook. I have a 233 tillamook that isn’t modded that I plan on using to test the mods.

-The pads bridge kinda easily, I’m going to pull the solder mask right to the edge of the holes.
-Depending on how the testing goes it looks like I may just remove the power features from the tweaker entirely. But I’ll try to avoid it.

Socket footprint looks pretty good,
It installs pretty easily, !and is even removable! (with all the white circles soldered)

Here is a pic of the kapton tape which works as intended. Didn’t get a pic of. The tweaker sits flush and the solder secures it exactly as intended. So that part is working perfectly!

Looks great so far. Keep up the going work.

Looks like a successful install, all planes are still seperated. Gonna put some processors in there and make sure they don’t short

Edit:

Figured ai may as well solder the rest for a good test.

Ground Shorted to vcc3

So I’m gonna make some adjustments.

It looks like I can make the holes smaller.

Also pull the mask in all the way to the hole.

The install went better after some practice but I believe I can redesign the hole specs to make this much harder to happen now that I can see what’s going on in physical form.

I’m also strongly considering Removing the power taps entirely from the design except for minimal taps to make the mods work. It will take the points that need to be soldered (or rather can be soldered) from what you see now down to like 15? I’d have to count..Still debating that. Will do some more pokdering.

ended up canceling the latest version order and making a new one with what I learned from that old one I ordered long ago.
between the improvements since the version I ordered to today plus the improvements I've made with the info learned from this prototype I think the design is quite a bit more refined now.

I'm gonna probably make a few more minor tweaks to this before ordering but I think this is pretty representative of what the next prototype will look like. and will probably order soon.

I was able to get those two vcc2 pins hooked up given the new hole sizes and that also allowed more copper between the pads which is nice for more current.

For now, I'm still holding onto the power feature. i think depending on how this next prototype goes it will probably make that desicion.

Also discovered and fixed a minor misalignment of a row of pins at the top of the lower segment. I assume it's present on all previous versions that include these holes. fixed now. it was minor, it still fit though. More critical now that the holes are smaller.

down the road I may be able to use this as a base for a processor based pcb. these pcbs are thin enough that they should be able to fit between the socket and processor.

Though I’m not sure how you guys plan to solder them on. Many pins are not along perimeter.

It may be possible to do a interference fit though.

Obviously this would drop the power features. And be smaller.

Socket 5/7/SS7 (Processor Shim) Tweaker (Released)

update: I went ahead and made it 🤣

pretty basic design. only question is if the holes are correct size for interference fit