I thought I would drive in the issue with clearance around the socket a little more. This is just on the same motherboard, I have one near miss that would get hit with a tantalum on the bottom, and another direct hit. You can see a jumper directly hitting the PCB. If that jumper is under the VRM section, it will hit any component. So I think it is a good idea not to even leave solder pads for a bottom cap as a jumper will short on it.

Sphere478 wrote on 2022-12-07, 03:03:

Some of us make ramen in the coffee pot man. […]
Show full quote

Some of us make ramen in the coffee pot man.

feipoa wrote on 2022-12-07, 02:39:

Hmmm... did I say that somewhere? If so, can you point me to the typo and I can edit it. So far, there was no net average impro […]
Show full quote

Sphere478 wrote on 2022-12-07, 02:10:

So it sounds like waveform gets better the more decoupling is added all the way to 21?

Hmmm... did I say that somewhere? If so, can you point me to the typo and I can edit it. So far, there was no net average improvement of 21 caps over 10. I am going to take it down to 8 and check again. EDIT: ahhh, only slightly better on a per-pin basis, but not as an average. EDIT2: if you want even better per-pin, wait for the results and discussion.

Sphere478 wrote on 2022-12-07, 02:10:

Anything I am missing? Anything else? :-)

I like what rasz_pl did with the decoupling caps. If not possible, forget about it.

I have a few more days of testing.

Yesturday's photos coming soon...

Sorry man, there is a lot of info in this thread and a lot of posts I’m tryin to keep up, doin a thousand things here.

If I forget or mis quote just remind me :-)

I’m searching posts for what rasz_pl did with decoupling caps. I forgot what they did? Do you mean with the flat on one side round on other pads? I know that will cause issues on some of the locations, but let me see if it could work for 8 locations?🤔

I see where he wanted me to pull back the 5v plane. Might work, but I feel iffy about pulling from so few pins. I also see more possible problems from that and no real benefit. What if mobo didn’t route those well?

Gonna try and make a better marking for pin one like requested

Pulling copper away from decoupling cap/socket pins may be less of an issue than square pads. But also has issues of reducing cross section.

I’ll stand by for your updates and conclusions.

If you are going to try it with 8 pins, I recommend these for location and a symmetric look.

Okay, lemme look at those locations 😀

This might be doable in a few spots.

This size and shape can be used without hurting the cross section in a number of places.

The round holes and oval pads of the regulator are messing up the clearance over rides. I believe I already solved this problem though. So I’m going to just leave it. The other option is to change the global clearance of that flood. 🤔 which I think is unwarranted at this time as I think I have already solved the problem.

These are pretty even and unobstructed. Approve?

Sphere478 wrote on 2022-12-07, 04:06:

These are pretty even and unobstructed. Approve?

Yeah, if that's what worked out conveniently. Are you able to make the pad the size rasz showed in my above image?

The image you provided is looking at GND and OUT, but the issue I'm having is IN and GND. See high-res image here:

See how the two triangles make an hourglass with the 5V plane (2nd pin from left)? Are you able to get this side of the 5V plane further away from GND (middle pin)?

Able, yes. But about half the locations would cut into cross section. I think the pad size I put in is gonna work. Remember we are soldering to the pin not the board.

Btw, hey rasz_pl.

How did you make that shape? Two overlapping pads? That’s what I did. It generates a drc and a overlap into the hole on 3d viewer 🤣. Is there a better way?

feipoa wrote on 2022-12-07, 04:39:

The image you provided is looking at GND and OUT, but the issue I'm having is IN and GND. See high-res image here: […]
Show full quote

The image you provided is looking at GND and OUT, but the issue I'm having is IN and GND. See high-res image here:

See how the two triangles make an hourglass with the 5V plane (2nd pin from left)? Are you able to get this side of the 5V plane further away from GND (middle pin)?

VRM_GND_problem_zone_1.JPG
VRM_GND_problem_zone_2.JPG

😀 also, silk there. All fixed! ..?

Latest Silk for reference

🤔 lookin good.

Figured out how to improve the cin and cout vias to a ridiculous amount. And took the time to make a custom flood for them.

Thoughts?

I think this should really be nice.

I’m sure there will be a few more tweaks over the next few days. But are there any concerns I didn’t address?

saving progress. feel free to look. not a release.

Were you able to make the 8 solder pads like this? I cannot tell from the photos you provided, but looks like you are content with solder mask? Doing it like shown below will reduce goof ups during assembly.

Normally on these adaptors, I see a big white triangle to identify pin 1. Inserting 386/486 incorrectly is a very common problem, so might want to consider:

On the bottom, I'd be a little more direct and write something like:
for PGA-168 SXL/SXL2 CPUs only

Without the only, one may think, 'ok, works with SXL chips, but maybe works with others too. Let's try... F'

Didn’t do as close together, but yes I did do that. See the green outlines screen grab for shape of pads now download/file.php?id=151998&mode=view

I marked pin one better, but you REALLY wanted it marked haha. How’s this? 😀

Sphere478 wrote on 2022-12-07, 06:44:

Didn’t do as close together, but yes I did do that. See the green outlines screen grab for shape of pads now download/file.php?id=151998&mode=view

I marked pin one better, but you REALLY wanted it marked haha. How’s this? :-)

Yes, the world will thank you for that white triangle. It will likely save a CPU or two in the years to come.

I've finished up my testing. I ran the scale at 100 mV this time and did at least a dozen perturbations. Will present later, some changes needed to PCB to make the MIC perform better, closer to the LP. Broader spectrum of VRMs supported, I think the better, especially will all these recent 'out of stock' component listings on mouser/digikey.

Yeah, availability of regulators is one of the primary reasons I decided to just go external power supply on my socket 5/7/ss7 voltage interposer tweaker. Also with target voltages between 0.3 and 1.3v lower than available supply voltage it seemed iffy to try and regulate from available on board supply, so external power source was needed anyway. May as well be a in line power adapter. Aslo linear isn’t a good option after socket 3 really. So complexity was higher and different for each option if one went nla Socket 5/7 technically has two vcc5 pins but I believe many later boards didn’r route them. So powering from that was out Also only two pins for vcc5 is sus. Sorry tangent 🤣

Anyway. Yea, you should give some thought to future proofing this. I can make a special version with terminal pads for external power pretty easily. Just replace the regulator and remove the trimmer and resistor traces/pad And include it in release as a alternate for future retro goers

Coming soon... This was the request of someone who donated to me.
Results within the next few days. The rest in a video and then to the store.

Ooo a nice side by side design! 😀

Hey Feiopa, whatcha think of using these?

😂

feipoa wrote on 2022-12-07, 03:34:

Removing those 4 tantalums from the centre region, which were 22 uF, 470 nF, 150 nF, 220 nF:
Noise on Vcc5 mostly disappears, or down to 48 mV and the noise on Vcc3 remains about the same.

fascinating, so they just acted like antenna?

Sphere478 wrote on 2022-12-07, 03:57:

This size and shape can be used without hurting the cross section in a number of places.

The round holes and oval pads of the regulator are messing up the clearance over rides. I believe I already solved this problem though. So I’m going to just leave it. The other option is to change the global clearance of that flood. 🤔 which I think is unwarranted at this time as I think I have already solved the problem.

moar, pull more copper away
click on pad, E, overrides, enter 0.4, B to refill

Sphere478 wrote on 2022-12-07, 04:41:

hey rasz_pl.
How did you make that shape? Two overlapping pads? That’s what I did. It generates a drc and a overlap into the hole on 3d viewer 🤣. Is there a better way?

+ manual 0.4 clearance

whats with all the silk screen over the solder pads?

Okay, I’ll look some more later 😀

Silk is clearanced in gerbers when kicad generates them.

The silk is extra protection against these shorting problems.

I think it makes the decoupling caps look cool also. 😀

sphere 2 CPUInterposer386-486.zip:
- can cut away 1mm on three sides (left right and bottom) and still fit memw
- can even cut upper edge 1mm. pull vrmsense track lower in middle layer under J1 footprint, same for NIO and SO tracks, dont route them on the edge.
- moooaaarr clearance
- clearance is better than silkscreen