VOGONS


First post, by H.W.Necromancer

User metadata
Rank Newbie
Rank
Newbie

Please HELP,
I have a super annoying issue with this 775 board. Nothing fancy but it has DDR3 and I have some use for it. However the board works 100% OK only with Pentium 4, or Pentium D. However if I try anything later - Dual-core, C2D, C2Q it refuses to POST at all. Black screen, no post. Does not matter what bios is there, if there is a PCI-E VGA or not, ot what kind of memory you use. I have tried like 6-8 CPUs with no avail.
Thank you

Attachments

Reply 1 of 13, by TrashPanda

User metadata
Rank l33t
Rank
l33t

Odd the CPU support list has it supporting pretty much everything, except for some of the later Core2 extreme CPUs but that was likely fixed in a bios update or not at all due to the VRM.

https://asrock.com/mb/intel/g41m-vs3/#CPU

Have you checked the socket pins to make sure none are bent or damaged ?

All it would take is for one of the VCC pins Core2 uses to not be making contact for the system to not post correctly, other than that I can see no reason it wouldn't work.

Oh noes, the cap let the shmooo out 😁

Reply 2 of 13, by Horun

User metadata
Rank l33t
Rank
l33t

Agree look over the socket very carefully. If 800Mhz Pentium D work then any 800Mhz Dual core/Core 2 should also work.
Maybe something up when clocked above 800Mhz, failing Vreg caps can cause issues with faster core speeds...

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor.

Reply 3 of 13, by TrashPanda

User metadata
Rank l33t
Rank
l33t
Horun wrote on 2022-08-13, 19:05:

Agree look over the socket very carefully. If 800Mhz Pentium D work then any 800Mhz Dual core/Core 2 should also work.
Maybe something up when clocked above 800Mhz, failing Vreg caps can cause issues with faster core speeds...

Not necessarily, Core2 uses different VCC pins in the socket to Pentium4, if one or more of the VCC pins is corroded or bent and cant make contact correctly then that would cause some CPUs to work and others not not work at all due to bad voltage. Same for bad caps in the VRM or even a weak phase, these older boards had pretty terrible VRMs to begin with.

I would hit the CPU socket with contact cleaner along with the ram slots and see if that makes a difference.

Oh noes, the cap let the shmooo out 😁

Reply 4 of 13, by H.W.Necromancer

User metadata
Rank Newbie
Rank
Newbie
TrashPanda wrote on 2022-08-14, 01:24:
Horun wrote on 2022-08-13, 19:05:

Agree look over the socket very carefully. If 800Mhz Pentium D work then any 800Mhz Dual core/Core 2 should also work.
Maybe something up when clocked above 800Mhz, failing Vreg caps can cause issues with faster core speeds...

Not necessarily, Core2 uses different VCC pins in the socket to Pentium4, if one or more of the VCC pins is corroded or bent and cant make contact correctly then that would cause some CPUs to work and others not not work at all due to bad voltage. Same for bad caps in the VRM or even a weak phase, these older boards had pretty terrible VRMs to begin with.

I would hit the CPU socket with contact cleaner along with the ram slots and see if that makes a difference.

Yeah, it is a super cheapo board - I have seen a lot of failure on these on YouTube, but 1) different failures 2) in Hindi 😂 Anyway, there is one slightly bent pin, but the head is ok and alligned and probably it is not even crucial - I remember that some mont ago I cover this pin on C2D with some masking tape and put it in to my testing board and it posted. The other pins are as new.
VRM - I measured on drain of all the mosfet. Is that correct? (I am learning). There are 3 pairs of FETs - I am getting 12V on the DRAIN of one and 1.1-1.3V on the second in the pair. If the P4 is running in the board, there is 1.3V. If another not-running CPU the voltage there is like 1.1V, which looks like it is trying to start but cant identify the CPU and set the voltage.
--
It has nothing to do with the fact the CPU is 800Mhz Bus or more. I have handful of Dual-core and C2D and all behaves the same. The CPU is pretty cold because the board is not able to set it and start it. It leads me to a conclusion the balls under the socket are screwed up and some pins are not working thus the board cant know the cpu is in...
--
Yes, VRM are super shitty but it looks they are giving power.

Reply 5 of 13, by mockingbird

User metadata
Rank Oldbie
Rank
Oldbie
H.W.Necromancer wrote on 2022-08-14, 09:36:

It leads me to a conclusion the balls under the socket are screwed up and some pins are not working thus the board cant know the cpu is in...
--

If this board is like my ASRock P43C-ME from that era, the socket will be through hole... Uncommon, yes, but nevertheless true. My board also 'stopped' working after I preemptively re-capped it... I ought to test it with an old P4 to see if the symptom is the same as yours.

jw0vu3.png
(Decommissioned:)
7ivtic.png

Reply 6 of 13, by Towncivilian

User metadata
Rank Newbie
Rank
Newbie

Is the BIOS the latest on this motherboard? I wonder if it is worth checking whether the appropriate microcodes are present in the BIOS using e.g. Microcode Extractor: https://github.com/platomav/MCExtractor

HP d530 CMT: P4 EE 3.2GHz, 4GB DDR-400, ATI Radeon 9600XT, SB X-Fi XtremeMusic, WD VelociRaptor 80GB, WD Caviar Blue 500GB

Reply 8 of 13, by H.W.Necromancer

User metadata
Rank Newbie
Rank
Newbie
Roman555 wrote on 2022-08-14, 20:14:

Have you tried to clear CMOS before changing a CPU ?

Yes, I had tried this.
As there is nothing to loose I am thinking of a reflow of the chipset and/or the LGA 775. However I need to wait for a bottom preheater to prevent the LGA from being melted.
If no avail I will use the board as a part donor for something more interesting.
--
There needs to be some pin/signal that Dual-Core and higher needs and P4 does not...It would be interesting to find out...

Reply 9 of 13, by Roman555

User metadata
Rank Member
Rank
Member
H.W.Necromancer wrote on 2022-08-15, 09:35:

There needs to be some pin/signal that Dual-Core and higher needs and P4 does not...It would be interesting to find out...

RT8841 PWM controller datasheet
A page 17 of the datasheet: Pentium 4 and Core 2 Duo have different version of VRM .
So check the VIDSEL pin and its joined components.

[ MS6168/PII-350/YMF754/98SE ]
[ 775i65G/E5500/9800Pro/Vortex2/ME ]

Reply 10 of 13, by H.W.Necromancer

User metadata
Rank Newbie
Rank
Newbie
Roman555 wrote on 2022-08-15, 16:15:
RT8841 PWM controller datasheet A page 17 of the datasheet: Pentium 4 and Core 2 Duo have different version of VRM . So check t […]
Show full quote
H.W.Necromancer wrote on 2022-08-15, 09:35:

There needs to be some pin/signal that Dual-Core and higher needs and P4 does not...It would be interesting to find out...

RT8841 PWM controller datasheet
A page 17 of the datasheet: Pentium 4 and Core 2 Duo have different version of VRM .
So check the VIDSEL pin and its joined components.

Thank you! If I understand correctly the CPU is connected to to VID 0-7 PINS in order to set the right voltage by pulling those low / high according the voltage tables?
But what does the VIDSEL pin? I mean how to check, what should i measure?
As I the mosfets look like there are all ouputting power whern P4 is there, I assume this:
1)either the RT8841 is half-faulty, 2)or there are some connections messed up under the socket which leads to improper CPU voltage detection. (?)
So if I drop some C2D in to the socket and measure the VID pins, I should get readings coresponding to the voltage tabel?
And what should I find on VIDSEL if the IC is working right?
Thank you very much!

Reply 11 of 13, by Roman555

User metadata
Rank Member
Rank
Member
H.W.Necromancer wrote on 2022-08-17, 09:16:
Thank you! If I understand correctly the CPU is connected to to VID 0-7 PINS in order to set the right voltage by pulling those […]
Show full quote

Thank you! If I understand correctly the CPU is connected to to VID 0-7 PINS in order to set the right voltage by pulling those low / high according the voltage tables?
But what does the VIDSEL pin? I mean how to check, what should i measure?
As I the mosfets look like there are all ouputting power whern P4 is there, I assume this:
1)either the RT8841 is half-faulty, 2)or there are some connections messed up under the socket which leads to improper CPU voltage detection. (?)
So if I drop some C2D in to the socket and measure the VID pins, I should get readings coresponding to the voltage tabel?
And what should I find on VIDSEL if the IC is working right?
Thank you very much!

Welcome! Yes.
As you can see on the page 17 (VID Table Selection) the RT8841 has 4 modes (2 for Intel's CPUs and 2 for AMD's CPUs).
VIDSEL is input pin of PWM and it is connected to a corresponding pin of CPU.
The signal selects a mode the PWM works.
Also VIDSEL can be connected to other components of the board.
Pentium 4 pulls the signal to GND (VR10.x mode)
Core 2 Duo pulls the signal to VTT (VTT = 1.2 V, VR11 mode) or just doesn't pull to GND (VTT appears on the VIDSEL pin via a pull-up resistor).
So (the board is powered-OFF and no CPU installed) first check using an ommeter if the VIDSEL is unnecessary shortwired to GND or not.
Then try to find other components that are wired to the VIDSEL. It would be much easier to measure voltage on that components then do it on the pin of PWM.
Measurement on a powered board is dangerous because a wrong action can be damage a board.
I think the RT8841 is good. It powers P4 well as you see the board running.
The measurements of VID pins are not necessary at this stage.

[ MS6168/PII-350/YMF754/98SE ]
[ 775i65G/E5500/9800Pro/Vortex2/ME ]

Reply 12 of 13, by H.W.Necromancer

User metadata
Rank Newbie
Rank
Newbie
Roman555 wrote on 2022-08-17, 16:56:
Welcome! Yes. As you can see on the page 17 (VID Table Selection) the RT8841 has 4 modes (2 for Intel's CPUs and 2 for AMD's CPU […]
Show full quote
H.W.Necromancer wrote on 2022-08-17, 09:16:
Thank you! If I understand correctly the CPU is connected to to VID 0-7 PINS in order to set the right voltage by pulling those […]
Show full quote

Thank you! If I understand correctly the CPU is connected to to VID 0-7 PINS in order to set the right voltage by pulling those low / high according the voltage tables?
But what does the VIDSEL pin? I mean how to check, what should i measure?
As I the mosfets look like there are all ouputting power whern P4 is there, I assume this:
1)either the RT8841 is half-faulty, 2)or there are some connections messed up under the socket which leads to improper CPU voltage detection. (?)
So if I drop some C2D in to the socket and measure the VID pins, I should get readings coresponding to the voltage tabel?
And what should I find on VIDSEL if the IC is working right?
Thank you very much!

Welcome! Yes.
As you can see on the page 17 (VID Table Selection) the RT8841 has 4 modes (2 for Intel's CPUs and 2 for AMD's CPUs).
VIDSEL is input pin of PWM and it is connected to a corresponding pin of CPU.
The signal selects a mode the PWM works.
Also VIDSEL can be connected to other components of the board.
Pentium 4 pulls the signal to GND (VR10.x mode)
Core 2 Duo pulls the signal to VTT (VTT = 1.2 V, VR11 mode) or just doesn't pull to GND (VTT appears on the VIDSEL pin via a pull-up resistor).
So (the board is powered-OFF and no CPU installed) first check using an ommeter if the VIDSEL is unnecessary shortwired to GND or not.
Then try to find other components that are wired to the VIDSEL. It would be much easier to measure voltage on that components then do it on the pin of PWM.
Measurement on a powered board is dangerous because a wrong action can be damage a board.
I think the RT8841 is good. It powers P4 well as you see the board running.
The measurements of VID pins are not necessary at this stage.

Thank you very much! This very good info from you! That is why I am doing that. The board is more or less worthless, but I wish to learn something about VRMs. Have a nice weekend.

Reply 13 of 13, by Roman555

User metadata
Rank Member
Rank
Member
H.W.Necromancer wrote on 2022-08-20, 11:56:

Thank you very much! This very good info from you! That is why I am doing that. The board is more or less worthless, but I wish to learn something about VRMs. Have a nice weekend.

Thanks!
A schematics (for free) for a real mainboard (but another one) with RT8841
It might be helpful.

[ MS6168/PII-350/YMF754/98SE ]
[ 775i65G/E5500/9800Pro/Vortex2/ME ]