Doornkaat wrote on 2021-04-08, 18:40:

That's 100% the hotfix that Gigabyte supposedly inofficially suggested. What's also weird is imho that the PBC has discoloration […]
Show full quote

That's 100% the hotfix that Gigabyte supposedly inofficially suggested.
What's also weird is imho that the PBC has discolorations at the other transistor but not the one supposedly responsible for +3.3V on AGP.
Could it be the transistor is supposed to only generate 1.5V and the other one is responsible for 3.3V?
Could you try starting the board and measuring voltages of both transistors at the back?

Edit: Reding the comments here this seems to be the case. The linked picture is wrong. I'll edit that post.

Interesting.

What I have now tried is removing the LDO module, and connecting as the picture shows. I now have 3.3v in the bios VGTL (as you'd expect).

You are right though, I noticed the "other" LDO module was getting warm when the board was in bios mode (I have this board just on a desk nothing attached bar the video card), but the LDO in question was not. I found this odd, as I was obviously drawing power from the AGP bus (the luke warm LDO module) and even though nothing was in the PCI slots (I assumed this was what the "warm"module was supplying) this was getting warmer.

With no modded connection, and no LDO module connected (removed) I still had approx 0.5 volt there.

I would like to get to the bottom of this, ideally replacing the underpowered LDO module with something of the correct Current rating (7A?)

But as of now things seem on the face of it, better.

Just as an aside, I've noticed if I connect the PC speaker of have constant high/low beeping, indicating a CPU issue. But it boots into bios OK. In hardware monitoring the -12 line is showing -62.3...Odd! EDIT!! The beeping goes away with an older PSU with a -5v(white) line, and shows the correct -5v in the Bios, so I'm assuming this Mobo doesn't like "newer" PSU's)!!

Also I can only set the FSB to 66mhz to boot. 100 gets the fans running but nothing more.

I'm really not sure what the 1.5v line would be used for though?

To be continued!

First: I understand with the mod wire connected to the right pad it works? If so, awesome! 😁

With the rest I'm getting a bit confused. Let's call the regulators LDO1.5 and LDO3.3, ok?

So which one is getting warm?

AGP2.0 specifies Imax = 8.0A for both Vddq1.5 and Vddq3.3. Maybe to comply to AGP2.0 Gigabyte added a 1.5V source even though the board does not support 1.5V signalling? Admittedly not convincing but maybe a place to start investigating.
Maybe AGP pin A2 somehow triggers LDO1.5 to supply the slot instead of LDO3.3 when pin A2 is connected to ground (indicating a card using 1.5V signalling)?

Anyway, if you want to use the onboard regulator instead of the wire mod AGP2.0 specifies Imax of 6A on Vcc3.3. What we do not know is wether the board also uses the regulator to create Vddq3.3 which Intel predicts to be realistically below 2A.
All in all assuming the AGP2.0 spec is our reference worst case the regulator would have to supply 14A @3.3V. With the Voodoo3 3000 something around 6A combined current draw for all 3,3V rails on the AGP slot appears more realistic to me but again that's just an educated guess!
Honestly I'd just keep the wire mod in place if it proved working. Even if only AGP Vcc 3.3V is provided by that regulator you're looking at 4-5A power draw for the Voodoo3 and that at least warrants a heatsink on the regulator.

Doornkaat wrote on 2021-04-09, 10:27:

First: I understand with the mod wire connected to the right pad it works? If so, awesome! :D […]
Show full quote

First: I understand with the mod wire connected to the right pad it works? If so, awesome! 😁

With the rest I'm getting a bit confused. Let's call the regulators LDO1.5 and LDO3.3, ok?

So which one is getting warm?

AGP2.0 specifies Imax = 8.0A for both Vddq1.5 and Vddq3.3. Maybe to comply to AGP2.0 Gigabyte added a 1.5V source even though the board does not support 1.5V signalling? Admittedly not convincing but maybe a place to start investigating.
Maybe AGP pin A2 somehow triggers LDO1.5 to supply the slot instead of LDO3.3 when pin A2 is connected to ground (indicating a card using 1.5V signalling)?

Anyway, if you want to use the onboard regulator instead of the wire mod AGP2.0 specifies Imax of 6A on Vcc3.3. What we do not know is wether the board also uses the regulator to create Vddq3.3 which Intel predicts to be realistically below 2A.
All in all assuming the AGP2.0 spec is our reference worst case the regulator would have to supply 14A @3.3V. With the Voodoo3 3000 something around 6A combined current draw for all 3,3V rails on the AGP slot appears more realistic to me but again that's just an educated guess!
Honestly I'd just keep the wire mod in place if it proved working. Even if only AGP Vcc 3.3V is provided by that regulator you're looking at 4-5A power draw for the Voodoo3 and that at least warrants a heatsink on the regulator.

The LDO I have removed is the "LDO 1.5".

Is the VGTL the AGP voltage? I can find no reference to this Online.

As I said the warmer LDO was the "3.3v" one, which confused me, as if it was supplying the PCI bus, this was empty, ie no current.

I've attached pictures

Many thanks

There are still devices on the pci bus even with no cards installed. The onboard ide controller is on the bus and drawing power and so are the various on board devices as well as the pci-isa bridge.

chrismeyer6 wrote on 2021-04-09, 12:49:

There are still devices on the pci bus even with no cards installed. The onboard ide controller is on the bus and drawing power and so are the various on board devices as well as the pci-isa bridge.

The only thing plugged into the board was CPU, RAM, and GPU in the AGP slot. No HDD's or PCI/ISA cards

I understand that but there are pci and Isa devices built into the motherboard like your ide controller that all draw power. Not everything is a pci or Isa expansion card.

Right, I've just tried a PCI GPU in a PCI slot. It boots into bios fine with the Voltage mod (and no "1.5V" LDO) - BUT, if I remove the mod (No LDO or VMOD) the machine doesn't boot. Seems to point towards it supplying power to the PCI bus/too?

chrismeyer6 wrote on 2021-04-09, 13:00:

I understand that but there are pci and Isa devices built into the motherboard like your ide controller that all draw power. Not everything is a pci or Isa expansion card.

Yes I understand, but enough to make it hotter than the LDO with a GPU attached?

It's very possible. It's probably supplying power to the ide controller, NB,SB, pci-isa bridge ECT. Without a schematic we don't know exactly what components are drawing off of it.

chrismeyer6 wrote on 2021-04-09, 13:08:

It's very possible. It's probably supplying power to the ide controller, NB,SB, pci-isa bridge ECT. Without a schematic we don't know exactly what components are drawing off of it.

Yes, maybe you're right.

I'm guessing the only way to know for sure, is, if someone who has the same Motherboard could check the output and control leg voltage. 🙁

Right after more research, and after Doornkaats reply, I have restored the 1.5v supply on the board. I have ordered a "LT1584CT T0 220" LDO, which will arrive in a few days. This is rated @7A, so I am going to replace the original 3.3 (next to the PCI top slot) with this (might have to alter the resistance to get the correct output voltage) and put a heat sink on it to see what happens. (I would prefer to add another voltage regulator just as a fail-safe if the PSU throws a wobbly!

I'll let you know how I get on.

Sounds good. I hope you can get this sorted out and start enjoying the system.

As some have mentioned, the posted links are wrong. That is a 1.5V regulator. The other one is the 3.3V where the wire should be connected.