Is anyone making these guys to sell? I'd love to pick a couple up. 😀

I would also like to buy one.

Me too.

Don't waste your time. The switching power module they are designed around are crap and can't maintain voltage under load.

kingcake wrote on 2024-06-30, 21:43:

Don't waste your time. The switching power module they are designed around are crap and can't maintain voltage under load.

How do you know?
Are there better solutions?

Kingcake's reply is about earlier revisions that used a cheap regulator module. Calamitylime's 4 layer redesign shouldn't have any of these issues. Unfortunately this redesign may not be finished yet. As far as I know it is yet to be tested.

kingcake wrote on 2024-06-30, 21:43:

Don't waste your time. The switching power module they are designed around are crap and can't maintain voltage under load.

Would running two modules in parallel potentially alleviate that (distributing the load) ?

akimmet wrote on 2024-08-19, 21:57:

Kingcake's reply is about earlier revisions that used a cheap regulator module. Calamitylime's 4 layer redesign shouldn't have any of these issues. Unfortunately this redesign may not be finished yet. As far as I know it is yet to be tested.

That redesign looks interesting, but there haven't been any updates in more than 1.5 years. Seems like it was abandoned?

darry wrote on 2024-08-19, 22:31:

kingcake wrote on 2024-06-30, 21:43:

Don't waste your time. The switching power module they are designed around are crap and can't maintain voltage under load.

Would running two modules in parallel potentially alleviate that (distributing the load) ?

No, those regulator modules just have extremely poor performance.
Connecting two voltage regulators in parallel is generally considered a bad idea. Unless they are perfectly matched, one of the regulators will end up doing all of the work. Designs that do use regulators in parallel, have balance resistors in series with each of the regulator outputs. These balance resistors hurt regulation performance.

TheMobRules wrote on 2024-08-19, 23:38:

akimmet wrote on 2024-08-19, 21:57:

Kingcake's reply is about earlier revisions that used a cheap regulator module. Calamitylime's 4 layer redesign shouldn't have any of these issues. Unfortunately this redesign may not be finished yet. As far as I know it is yet to be tested.

That redesign looks interesting, but there haven't been any updates in more than 1.5 years. Seems like it was abandoned?

I've been very tempted to try this project out, but I have far too many projects on my workbench as it is.

Calamitylime made a external power supply board for my interposers. Perhaps his external power supply would work with this socket interposer? If you could find an appropriate place to inject the power.

akimmet wrote on 2024-08-20, 16:40:

darry wrote on 2024-08-19, 22:31:

kingcake wrote on 2024-06-30, 21:43:

Don't waste your time. The switching power module they are designed around are crap and can't maintain voltage under load.

Would running two modules in parallel potentially alleviate that (distributing the load) ?

No, those regulator modules just have extremely poor performance.
Connecting two voltage regulators in parallel is generally considered a bad idea. Unless they are perfectly matched, one of the regulators will end up doing all of the work. Designs that do use regulators in parallel, have balance resistors in series with each of the regulator outputs. These balance resistors hurt regulation performance.

Not all regulators are the same, some use a constant current source instead of a voltage references. Those can be easily paralleled. The LT3080 is like this, and examples of paralleling them can be found in the datasheet. They do still recommend some 50mOHM balance traces. (You don't need resistors, a squiggly trace on the PCB is enough to balance them).

You can also just use an op-amp to monitor the current output to keep them balanced. This applies to almost all regulators.

Sphere478 wrote on 2024-08-20, 23:06:

Calamitylime made a external power supply board for my interposers. Perhaps his external power supply would work with this socket interposer? If you could find an appropriate place to inject the power.

I just used a Linear Tech regulator specifically designed for 3.3V CPUs and made a supply board the same size as the switcher originally specced. So I can whack it on pre-made socket blasters. Easy and cheap solution.

I got the 486SocketBlaster to work.
Have been running doom and quake in demo-mode for about 30 min. Looks stable. Only with a heatsink - no fan.
My Am 486 DX4-100 is actually an Am 5x86 133 if i put the Jumper to 4X mode.
This is the Motherboard i am using: https://stason.org/TULARC/pc/motherboards/E/E … 86-UM-486V.html
More info coming soon.

Update:
This is the full story. First I ordered 10 PCBs from PCB Way. I soldered one of them. I did not get it working. I then realized that have to do a better job at cleaning and making sure i did not create a short circuit between the pins. I ended upp giving away 6 PCB's and necessary components, which i bought Pins and Voltage Regulator from AliExpress, and capacitors at Elfadistrelec.
Of the 6 PCB's i gave away i know that one of them is soldered and working.

I then ordered 5 fulle assembled 486SocketBlaster from PCW Way. I had a few rounds with pictures back and fourth, before they understood how important that the pins was soldered exact.
When i received them i saw that the Voltage Regulator was soldered to high, so i had to resolder them, to avoid hitting the underside of the CPU.
Of the Five, two of them did not work. Two works fine. And one I am not sure yet.
The problem I have with the third is that when I set the Voltage to 3.3, i get 4,6V after inserting the Intel DX4-75. I could remove one of my AMD CPU's from their 486SocketBlaster to test, but I dont want to risk damaging the Sockets. I have ordered a AM 486DX4-100 from eBay, so i will try with that when it arrives.

My thoughts on bying them assembled from PCB Way.
I really had no hope that PCB Way would be able to solder this, especially the pins. But they managed to get 2 (maybe 3) out of 5 working. This was way better than i had thought. I took a chance, and it paid of. If I get the third working, I have alle the 486SocketBlaster i need for now.

Nice!

Here are some pictures from using the 486SocketBlaster with a AM 486DX2-66. I am running it at 100Mhz.

Here is my experience with the making, and buying 486SocketBlaster https://youtu.be/YV4V_aUYyPI

Jinxter wrote on 2024-09-30, 16:11:

Here is my experience with the making, and buying 486SocketBlaster https://youtu.be/YV4V_aUYyPI

couple notes:

- on your pictures of 2 not working sockets one has regulator soldered wrong way = smoke
- you said you didnt like them sticking out so you resoldered it flush by pushing down. This made regulator short to the socket PCB = passing input 5V to output

Im impressed by pcbway managing to assemble something this custom, and even fix up after your notes.

rasz_pl wrote on 2024-09-30, 23:27:

couple notes:
- on your pictures of 2 not working sockets one has regulator soldered wrong way = smoke
- you said you didnt like them sticking out so you resoldered it flush by pushing down. This made regulator short to the socket PCB = passing input 5V to output

You have a keen eye. The VR that is thew wrong way, is the one that smoked, i took it out and replaced. So what you see it the smoked VR just placed on top of the 486SB.
Will i dont mind them sticking out, but that would hindered the CPU to be seated all the way down. There is really nothing on the backside that can short. Or?

In that case you probably had bad batch of those bigger regulators 🙁 its possible for counterfeit chip to regulate fine free-wheeling, but stay shut (thus passing input) under load.