I feel like someone in this scene is more likely to have through hole parts kicking around vs SMD parts.

For example:
I have a few SMD bits kicking around but I have no idea if any of them are even resistors, never mind the correct value but I know I have 1k THT resistors or at least values close enough to 1k.
I think for boards that are small and hacky, through hole parts are the way to go.
Correct me if I'm wrong.

I didn't do any trace smoothing or teardrops, people can add them if they want in kicad I suppose.

O.K. I finally found out how to reach the contacts TIO 8, TIO 9 and TIO 10 of the HX Norhbridge when they are not connected to the mainboard!

Since the one on my Jetway board is broken anyway I desoldered it. First I found that, for whatever reason, the pad for TIO 10 - and ONLY this one - was connected to a trace going somewhere (will find out where). This needn´t be the case for different mainboard models.

Then I cleaned the pads on the chip and started the measuring.
Here´s the official Intel pinout (view from the top) I marked the three I/O-lines:

Then I measured the connectivity between the 3 ballpads and the tiny traces on the chip.
This is the result:

This would not have been possible without removing the BGA chip, so the Jetway HX mainboard has been sacrificed - for science, but now it´s possible to enlarge the TAG IO-bus for mainboards with just one TAG chip that provide no connections at all.

This is going to have to be documented somewhere.
Should this topic of the tag ram be it's own thread maybe?

Honestly I don't think I've seen a mod where you solder wires to the top of a BGA chip, so me that's quite cool. I think I'm going to have to try this out.
Great work man.

-Lime

We still have to test this on a real HX-mainboard that lacks the 2nd TAG chip. It would make things easier if the existent TAG-socket was near the northbridge. The traces on the chip (in fact it´s a very thin multi-layer PCB with the actual silicon somewhere inside at the center) are quite delicate, so only very flexible and thin wires can be used. After the soldering is done and when everything works, a drop of epoxy resin could be applied.

And there´s an explanation for the trace at TIO 10:

This means when we switch from standard 64Mb cacheable area to the 'extended" cacheability of 512M this pullup at TIO 10 must get interrupted (due to the exclusiveness). After cutting the trace the side leading to the northbridge can be used to solder the TIO 10 wire.

This is good news - in most cases there should already be a trace connected to TIO 10, we just have to find it - then there are just 2 connections left that need to be soldered to the chip.

majestyk wrote on 2022-05-31, 15:25:

This is good news - in most cases there should already be a trace connected to TIO 10, we just have to find it - then there are just 2 connections left that need to be soldered to the chip.

Well I poked around my board a bit and I think I found TIO 10. I'm not sure it makes sense where it goes though.
Following your image TIO10 is in the absolute corner of the chip, looking underneath my board it looks like there is a via for the corner most pin. Following the trace it lead to two of the ram slots.
Simm one and simm three

I included an image highlighting the pins.

I don't have something to poke under the chip to probe the ball, nothing I can think of anyway.

-Lime

Edit: I forgot to include the image

Edit: The pin number of the ram slot is 3 and it seems to be called DQ16, if the pinout I found is to be believed.
https://arstechnica.com/gadgets/2000/07/ram-guide-part1-1/7/

Just as well I went poking around the board.

I spotted a flake of solder under the chipset, I didn't rework that part of the board except to remove one cap that was in the way of a trace I thought was tio10 for a bit.
I have noticed that the board is missing two caps that I did not remove. So this board got some re-work before me. The two caps are part of the voltage regulation circuits.

Not sure what this really means for the board yet but I should replace those before powering it at least. The board uses only 2 values of caps so it's easy to guess the values of the missing caps.

-Lime

Don´t poke under the chip - too hard. Some solder-balls might lose contact. This green micro-pcb is quite flexible and can easily be bent.

On my (now FUBAR) Jetway J-5HXA the TIO 10 pin goes to an empty resistor position besides one of the L2 SRAMs. If the resistor was populated TIO 10 would have been pulled up, but it never was.

yeah, I was wrong
I took a look around and found some resistors whose legs were small enough. I don't know what pin I was looking at but the balls goes to a resistor right beside the chip.

I thought it was that resistor before but I assumed it would pull up to 5v and I didn't get a connection to 5v from it. Instead it went to the negative leg of a near by cap.

R138 is the resistor in question.
All I have to do is remove that and I have easy access to io10.

While I was at it, I filed away the mask of the traces you pointed out in a previous video. I'm willing to try soldering a wire to them.
Seeing the bit is rework lowed my expectations so who knows, could be fun.

Hello,

I decided to upload the exp slot cover and DS1287 Hat (renamed to DS128X Hat) to GitHub. Bare minimum upload and I could add pictures and stuff another time, just not in the zone for such things.

https://github.com/LimeProgramming/Expansion-slot-cover
https://github.com/LimeProgramming/DS128X_Hat

Three ds12885 dip chips arrived in the post today which is fun.

Hello,

I've ordered the tag ram modification boards plus two tag ram chips for tinkering with that project!

-lime

CalamityLime wrote on 2022-06-03, 07:56:

Hello,

I've ordered the tag ram modification boards plus two tag ram chips for tinkering with that project!

-lime

W00t!!

Btw, tyan s1564 has two tags if you need me to look at anything on it.

well can you tell us if pins 15,16,17,18 and 19 are pulled high or low on the second tag ram?

The board is pulling low since it's probably better than leaving them float but I'm curious what actual hardware does.

CalamityLime wrote on 2022-06-03, 08:10:

well can you tell us if pins 15,16,17,18 and 19 are pulled high or low on the second tag ram?

The board is pulling low since it's probably better than leaving them float but I'm curious what actual hardware does.

Which pins.

You want me to measure to gnd and to vcc3 or 5v?

https://www.ultimateretro.net/en/motherboards/5131

I highed the pins but i don't know which chip is tag 1 or tag 2.
I'm looking for tag 2.

I suppose if the pins are pulled to gnd or vss as you said. Could be a resistance of a few K on the line as well.
I doubt they would be wired in parallel but it's not an impossibly

-lime

Hold tough, I'm an idiot.

I marked the wrong side of the chip.

I went top pin to pin below and it’s 4 mega ohms. Therefore they aren’t connected to any pull ups or downs. except maybe logic inside the chips.

I did this on both chips

Sphere478 wrote on 2022-06-04, 00:03:

I went top pin to pin below and it’s 4 mega ohms. Therefore they aren’t connected to any pull ups or downs. except maybe logic inside the chips.

I did this on both chips

huh, interesting.
they are hardly in just parallel, are they?

CalamityLime wrote on 2022-06-04, 00:14:

Sphere478 wrote on 2022-06-04, 00:03:

I went top pin to pin below and it’s 4 mega ohms. Therefore they aren’t connected to any pull ups or downs. except maybe logic inside the chips.

I did this on both chips

huh, interesting.
they are hardly in just parallel, are they?

If both of those pins were pulled high or low I would have gotten something way below megaohm. If they were parallel I would have gotten 0 ohm.

I think you are probably right that those red circled ones are the stray pins as there are no traces visible on them in the pic.

I didn’t want to press my luck with a bunch of readings, the voltage from the meter on these chips makes me nervous.. even if I know it should be fine.

I figure that reading is enough to disprove pull up or down on those two pins.

Yeah, voltage meters in chips can be spooky.

Well thanks for checking what you checked.
Still, we'll see what comes of things whenever I get the PCBs . Though i still don't know if the motherboard works.
Yet to pick up a cpu for it.

I don't have a bíos code reader card so and i don't think they can beep without a cpu