Announcement

**socketa** · 11-14-2020, 05:13 PM

Re: ASUS P5GC-MX - No memory, no boot

I'v attached a close up of the MOSFETs on the bottom right corner, and and texted in the voltages that are present when the PSU is turned on.
Looks like they are not being switched on
The coil is attached to the drains of the two MOSFETs that are closest to it, so it's only got 0.048V on it

Also noticed that the capacitor that is attached to pin 4 (OUT) of AT9173 is dented inward - could that be causing the the chip to not turn on?
The two MOSFETS, that are on the right side of the RAM slots, have no significant voltages on them.

(Additional info)
There is a 3-pinned SMD component between the AT9173EG chip and the MOSFETS, that has 70 and 80 (at 90 degrees to the 70) printed on it - i can't find a datasheet for it. The outward pin is connected to pin 3 (REFEN) of the chip.

Attached Files

**momaka** · 11-14-2020, 10:05 PM

Re: ASUS P5GC-MX - No memory, no boot

Originally posted by socketa View Post

Draws 27W with just the hard drive connected.

Did you have RAM installed on the board when you tested this?

Being an LGA775 board, I suggest to remove the CPU and inspect the CPU socket for bent pins - especially the corner rows on all sides.

Also, looks like you got the better version of this board

- this is the P5GC-MX/1333, which I think supports even 1066 and 1333 MHz FSB CPUs (whereas the regular P5GC-MX only supports 533 and 800 MHz FSB.) And you got almost all Panasonic caps to boot (with the exception of the three KZG above the RAM... which can be suspect, but I'll give them a pass for now - I think they would have bulged if they were, as KZG pretty much always do when they fail.)

As for the voltage measurements you did... looks like the upper MOSFETs in the RAM VRM (for Vdd rail) are getting proper voltage (5V), but just not being switched on. Can't say yet if that's normal or not though. Generally, there should be 1.8-2V for the RAM, regardless if RAM is installed or not. But if the power-up sequence is held up for whatever reason, that may be why there is no voltage there, so it doesn't necessarily indicate a problem yet. And you really need to test with RAM to see the proper voltage.

**RJARRRPCGP** · 11-14-2020, 10:50 PM

Re: ASUS P5GC-MX - No memory, no boot

QFT, the socket is suspect! Knowing socket 775 too well!

**socketa** · 11-15-2020, 01:23 AM

Re: ASUS P5GC-MX - No memory, no boot

Did you have RAM installed on the board when you tested this?

No

I've now put two sticks of RAM in, and rechecked the voltages:
The only voltage that differs from the previous readings is the gate voltage on the MOSFET that is on the left hand side of the RAM slot DIMM A1, and that is now 10.7V.

I went to check the CPU, and noticed that the heatsink clamp on the bottom right was UP instead of down.
Removed the heatsink
Checked the pins (all look good and uniform)
Put on new thermal paste and reinstalled heatsink
Cleaned off black dust buildup from around the CPU MOSFETs etc

No change in symptoms/readings

Also checked the four MOSFETS that are above the CPU and in-between the two horizontally installed output coils
(in the first photo, you can see a coil and top two MOSFETS just above the heatsink fan, and the tip of the other coil a bit further along underneath the input coil)
GDS from left to right, top then bottom
1.3mV 12.07V 1.3mV
1.3mV 12.07V 1.3mV
0.7mV 1.3mV 0.7mv
0.7mV 1.3mV 0.7mV

So, they are not being switched on

And on the right hand side of the CPU there are two more MOSFETs and a lesser wound horizontal toroid
G : D : S
1.3mV : 12.07V : 1.3mV
0.7mV : 1.3mV : 0.7mV

So these other MOSFETS are not being switched on either

**ChaosLegionnaire** · 11-16-2020, 10:24 PM

Re: ASUS P5GC-MX - No memory, no boot

Originally posted by momaka View Post

Did you have RAM installed on the board when you tested this?

socketa said this board doesnt beep even without ram in the first post, therefore the problem must be from before ram initialisation of the post process. and now it looks like even the cpu vrm isnt starting up properly and supplying the right vcore voltage to start the cpu.

**momaka** · 11-19-2020, 06:08 PM

Re: ASUS P5GC-MX - No memory, no boot

Originally posted by socketa View Post

So these other MOSFETS are not being switched on either

Yeah, looks like the board is not even triggering its power-up sequence (or it is, but something is stopping it.) This could be either due to CPU not being detected or a problem with one of the regulated voltage rails.

Before anything else, though, try removing the CMOS battery and leave the board disconnected from PSU for about a minute. Don't install the CMOS battery back. Now plug in the PSU but do NOT try to turn on the motherboard. Check voltage on the 3.3V STBY rail. I'm not sure if this will be the same for the P5GC-MX/1333 (I only have the P5GC-MX, which appears as almost the same board, though not 100%), but I think your 3.3V STBY regulator should be that MOSFET between the PCI-E 16x and 1st PCI slot. IIRC, Drain on that MOSFET should be connected to 5VSB rail on the ATX connector and the Source pin is the output of the 3.3V STBY rail (or maybe it was a linear regulator and the voltages are the other way around, I don't remember exactly.) You should get 3.3V there all the time as long as the PSU is plugged in (doesn't matter if the motherboard is turned On or Off.)

On some motherboards (well, on Biostar and Shuttle, at least), when the CMOS is reset and board is plugged in to the PSU for the first time, the RAM V_dd/V_ddq and V_tt rails come On without the board being powered On. I don't remember if this motherboard was like that, but try checking RAM V_dd and V_tt, just in case to see if they can even have those values. Report back what you get for the 3.3V STBY, RAM V_dd, and RAM V_tt from these tests.

Next check...
Try removing RAM and CPU, then disconnect CPU. Now try measuring resistance to ground on CPU V_core, CPU V_tt, RAM V_dd, RAM V_tt, and chipset/NB Vcc rails. From these, only chipset/NB should show a low resistance to ground (probably 10-50 Ohms) and maybe RAM V_tt (might be anywhere from 10 to 200 Ohms.) The chipset/NB Vcc, again if I remember correctly, is generated by one of those MOSFET on the bottom side of the RAM slots. The other one is for the PCI-E 1.2V comm. bus and/or SB V_cc. Let me know what values you get for all of these.

Originally posted by socketa View Post

I've now put two sticks of RAM in, and rechecked the voltages:
The only voltage that differs from the previous readings is the gate voltage on the MOSFET that is on the left hand side of the RAM slot DIMM A1, and that is now 10.7V.

OK, I just wanted to check how the board would react with RAM in there.

Usually, a motherboard should beep an error code when RAM is not installed, but the RAM voltages (V_dd/V_ddq and V_tt) should still be present.

**socketa** · 11-21-2020, 01:31 PM

Re: ASUS P5GC-MX - No memory, no boot

Check voltage on the 3.3V STBY rail.

It's a Liner Regulator and the voltages are as you say - they are good

Report back what you get for the ... RAM V_dd, and RAM V_tt from these tests.

I cannot find a pinout diagram for Unbuffered DDR2 RAM that includes VDD, VDDQ, and VTT
From what i've seen, normal unbuffered DDR2 RAM doesn't have a VTT pin!
For example, here are three diagrams
These first two have no VTT pin
https://pinouts.ru/Memory/dimm_ddr2_pinout.shtml
https://allpinouts.org/pinouts/conne...-dimm-240-pin/
and the other (which is buffered, and i dont think is the same as mine) is:
https://manualzz.com/doc/2581541/tra...dimm-dual-rank
which has VTT pin; but no VDDQ.
Anyway, here is the voltage measurements (correlated to this pinout https://pinouts.ru/visual/gen/dimm_ddr2.jpg) for some pins that i think would include VTT and VDD
1 2 3 4
61 62 63 64
237 238 239 240
With no CMOS battery, and switched on at the wall:
0 0 0.035 0.035
0 0 0 0
0 0.005 0 0.005
With CMOS battery, and switched on at the wall:
0 0 0.0035 0.0035
0 0 0 0
0 0.005 0 0.005
With CMOS battery, and computer turned on:
0.024 0 0.005 0.005
0.005 0.048 0.005 0.48
0 3.3 0 3.3

Next check...
Try removing RAM and CPU, then disconnect CPU. Now try measuring resistance to ground on CPU V_core, CPU V_tt, RAM V_dd, RAM V_tt, and chipset/NB Vcc rails. From these, only chipset/NB should show a low resistance to ground (probably 10-50 Ohms) and maybe RAM V_tt (might be anywhere from 10 to 200 Ohms.) The chipset/NB Vcc, again if I remember correctly, is generated by one of those MOSFET on the bottom side of the RAM slots. The other one is for the PCI-E 1.2V comm. bus and/or SB V_cc. Let me know what values you get for all of these.

Now, looking at the diagram that has VDD an VTT (but no VDDQ), and assuming that it's pinned out in the same way as the other diagram
VDD = pin 1 and VTT = pin 237:

RAM VDD = 599 ohms
RAM VTT = short
CPU VCC = 20K ohms and increasing
CPU VTT = 153 ohms

As for the NB VCC:
there is a group of 4 paralelled MOSFETS below the ram slots, where two drains are connected to two sources which are connected to the toroid coil (D > S > Coil) :
403 ohms
and also a single one on the bottom right corner with a small coil and the drain is connected to the coil: 1409 ohms

(I measured CPU VTT on the CPU slot pin, as i didn't know where else to measure it)

**momaka** · 11-28-2020, 10:50 PM

Re: ASUS P5GC-MX - No memory, no boot

Give me a day or two, and I'll try to make you a diagram with the rails.

In general, though, that bigger toroid/inductor by the RAM slots is your RAM V_dd/V_ddq rail... or sometimes called V_dimm (depending on manufacturer, I guess.) It's the 1.8-2.1V that goes to the RAM. Then there is also another rail for the RAM (V_tt... which is usually 1/2 of V_dd/V_ddq, so about 0.9 to 1.0V) to keep all of the contents when the PC is in use or in standby. This RAM V_tt can be found on all of the small quad array SMD resistors by the right RAM slot (you'll see these SMD array resistors are tied to the same trace on one side - and that's the RAM V_tt). As for CPU V_tt, that's easy on this board - it's at the Source of the TO-252 MOSFET on the lower-left corner of the CPU heatsink / upper-left corner of the NB heatsink. Or, if we use this picture as a reference...
https://www.badcaps.net/forum/attach...1&d=1604982040
... that MOSFET is located in the upper-left corner of the CPU heatsink and upper-right corner of the NB heatsink. CPU V_tt should be around 1.2V to 1.3V for LGA775 CPUs, if I still remember correctly. Supply for that MOSFET should be the 3.3V rail from the ATX connector (and found on the Drain pin.)

As for the two MOSFETs under the RAM slots... one is for the NB Vcc voltage, and the other for SB, I think. But that's the part I don't remember well, so I'll have to check my own P5GC-MX. I have to service that system anyways - the last few KZG caps I left on it have finally hatched too. They aren't on any important rails though (two being USB port GP power filters, so no biggie.) Anyways, you won't find any significant voltage on these MOSFETs on your board (both on the Source or the Drain) if your RAM V_dd is not coming up.

**socketa** · 11-29-2020, 03:25 AM

Re: ASUS P5GC-MX - No memory, no boot

Cool, thanks
When i said "below the ram slots" and "bottom right corner", i was referring to first photo with the printing horizontal, (and it fits better on the monitor screen), so the ATX 24 pin connector would be on the bottom.
But it seems that the convention is to view the PCI slots as the bottom of the board - so i'll use that from now on.

I just tried it with another CPU (Pentium 4), but no change
No voltage on the RAM V_dd toroid, and no voltage on the common trace of the quad array resistors on right of the RAM slots

**piernov** · 11-29-2020, 08:40 AM

Re: ASUS P5GC-MX - No memory, no boot

You are lucky that both schematics and boardviews are available for this board (that's very rare on desktops even more Asus schematics), so use them.
Since you identified RAM VDDQ is missing, if there's no short to ground on it I'd start by checking voltages on POU1.

Attached Files

**SMDFlea** · 11-29-2020, 09:40 AM

Re: ASUS P5GC-MX - No memory, no boot

Looks like a trace is damaged too, unless its a dead insect or some other shit

Attached Files

**socketa** · 11-29-2020, 04:45 PM

Re: ASUS P5GC-MX - No memory, no boot

Looks like a trace is damaged too

The trace is good. Was just some dust that was causing that effect

You are lucky that both schematics and boardviews are available for this board

i cant view the files in the zip file (seems that they require Topsolid viewer to view them, so i installed a small version 6.12, but got an error message and the file wouldn't open. Is there some alternative lite software to view this? Or maybe a screenshot?)

Since you identified RAM VDDQ is missing, if there's no short to ground on it I'd start by checking voltages on POU1.

According to the RAM diagram on pinouts.ru, VDDQ is pin 62, and no significant voltage appears on it, and there's no short to ground (370 ohms with CPU installed and PSU plugged in)
i found POU1 in the PDF, but, unlike MSI boards, the components have no labels on the actual board; but i'm pretty sure that i've found it between the RAM slots and the CPU fan connector; as it's the same component in the schematic and there's only one instance of it

Attached Files

**socketa** · 11-29-2020, 06:28 PM

Re: ASUS P5GC-MX - No memory, no boot

The chip, POU1, is an APW7120KE, and it's a driver for MOSFETs
Pin 1 Boot 11.65V
Pin 2 UGate 0.02V
Pin 3 Ground 0V
Pin 4 LGate 0.02V
Pin 5 VCC 11.84V
Pin 6 FB 0.01V
Pin 7 OCSet 2.55V
Pin 8 Phase 0.01V

So no voltage on Pin 8, which i assume is the RAM V_dd/V_ddq

**piernov** · 11-29-2020, 10:27 PM

Re: ASUS P5GC-MX - No memory, no boot

Boardview can be opened with OpenBoardView. It'll help you find the components on the board.

Can you check diode mode measurement across POD1 and POD2? (red probe on pin 1 and black probe on pin 3)

**socketa** · 11-30-2020, 12:01 AM

Re: ASUS P5GC-MX - No memory, no boot

Ok,
Yeah, i tried OpenBoardView as well, and it didn't work on Windows 7 32 bit, so tried it with windows 7 64bit, and it worked
POD1 = 331 with one multimeter and 307 with another multimeter
POD2 = 306 with one multimeter and 285 with another multimeter

So that looks good(?),
and the "+12V" supply is going through them to the BOOT terminal on the driver chip (12V, minus the two diode voltage drops, to give about 11.65V)

**piernov** · 11-30-2020, 06:33 AM

Re: ASUS P5GC-MX - No memory, no boot

Yes that's ok, I was just confirming but in fact no reason for that to be the problem.
Not sure why the RAM power rail isn't created then, voltages look good but indeed no output. Maybe the buck controller itself is bad.

**socketa** · 12-01-2020, 01:41 AM

Re: ASUS P5GC-MX - No memory, no boot

In the first post i mentioned two AT9173 chips, and the datasheet says "The voltage output of the regulator can be used as a termination voltage for DDR SDRAM"
The REFEN pin of one of the chips is being pulled low (i think that's you you say it) and i think that that is causing the chip to shut down.

Someone posted some info on this thread:
http://********/viewtopic.php?f=224&t=24455&start=30
the site is vlab . su
(the link wont post properly)

Start Quote:

These boards have a trick ...
The microcircuit between the southern and the at9173eg memory slots flattens it in every possible way ...
If the voltage on the fourth leg is less than the working one (1.05V), then the board does not start, the percentage is heated, the South is abnormally cold.
If the voltage is more than 1.05V, then the board works but constantly hangs, the SOUTH heats up more than usual. in the picture 1.08 tupanul when I did the picture. on other Asus there are RT9173C which I install ... there is also RT9173B it is not desirable to install it (for the test you can) it is 1.5A.

And you are not mistaken again maybe 1.25, not 1.05, I have 1.06, according to datasheet 1.25

No, there is exactly 1.05 V. In short, I fucked with such a motherboard as the author had with the same problem, but only initially, on my LM324 on the 4th leg, the voltage was normal (something about 11.5 V). Prots also warmed up but on the 15th leg of the PCI slot - it was "0" V. That is, a stuck reset. I began to check the generation on quartz. There was no 25 MHz quartz at all as well as a network controller. There are only 2 quartz left - RTC and 14.3 MHz near the RTM876-660 clocker. So there was no generation at 14 MHz quartz and it was decided to replace it. And voila, the mother came to life!

End Quote

Unfortunately, these chips are not on the PDF, because the PDF is for a different version board

**piernov** · 12-01-2020, 08:48 AM

Re: ASUS P5GC-MX - No memory, no boot

The DDR power rail VDDQ is required before the termination voltage VTT can come up. VTT is referenced from VDDQ since VTT is half VDDQ. Also VTT is only required when actively talking to the RAM in S0 state. It's not required in S3 state when just refreshing it (while VDDQ is required to keep RAM content alive).

On the schematics the chip for DDR VTT is PWU4. It's a CM8562 not an AT9173 but it does the same job.
The other AT9173 on your board powers the ICH core well at 1.05V so that's fine since you measure 1.06V on it. Difference from the schematics here is that on your board it's derived from +5V, while on the schematics it's derived from +1.8V_DUAL and regulated by PWU1 (LM324) so in your case it's present even though RAM power rail is missing.

**socketa** · 12-04-2020, 11:01 PM

Re: ASUS P5GC-MX - No memory, no boot

OK, thanks for clearing that up.
So the next step would be to try and replace POU1?
Is it common for these to fail?

Announcement

ASUS P5GC-MX - No memory, no boot

ASUS P5GC-MX - No memory, no boot

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment