Intel DG965RY power on every time main AC power connected

[HexaPro]

Hi, i'm having problem with Intel DG965RY mainboard. It's always power on by itself for 2 seconds and then turn off every time i plug in the AC cord. It also had a buldging 6.3V 220uF KZG cap next to the north bridge which i have replaced with 10V 470uF nichicon from a donor board (there's three of them).

The mainboard would run normal but when it run Fuji printing program the process would take a long time (it is a PC dedicated for printing photos), the preview picture would appear very slowly in the program and the CPU seem to be working at full capacity while loading those image and it's not normal to be that slow. But the loading time seem normal when running other program like Photoshop or Nero. I have already reinstall the Windows and programs several times just to make sure but it still act the same when printing.

I've replaced the board with MSI P31 Neo which i have successfully recapped recently and it run the program a lot faster like it should be. So that means the Intel board has a problem, i just don't have a clue what could be the problem? Badcaps? Where? Which one? I haven't test the board again with that Fuji app after recapping though and it still power on by itself for 2 seconds when AC power plugged in.

[lti]

Re: Intel DG965RY power on every time main AC power connected

Replace all KZG caps. If there are any KZJs or Nichicon HM or HN made between 2001 and 2004 (the date code on Nichicon caps is a letter, the last two digits of the year, and the week number - H0425 would indicate the manufacture date as week 25 of 2004), replace them as well.

[momaka]

Re: Intel DG965RY power on every time main AC power connected

Hi, i'm having problem with Intel DG965RY mainboard. It's always power on by itself for 2 seconds and then turn off every time i plug in the AC cord.

I *think* that's a Intel-specific thing and not really a problem. Reason I say this is because I have seen many Dells with Intel boards that do this. My Dell Dimension does the same thing when AC cord is disconnected and then reconnected.
I also have a Gateway Select 750 that does this, and its motherboard doesn't have a single electrolytic capacitor (it's all tantalum and ceramic capacitors).

I'm not sure about the other issue, though.

[HexaPro]

Re: Intel DG965RY power on every time main AC power connected

Replace all KZG caps. If there are any KZJs or Nichicon HM or HN made between 2001 and 2004 (the date code on Nichicon caps is a letter, the last two digits of the year, and the week number - H0425 would indicate the manufacture date as week 25 of 2004), replace them as well.

Thanks, i'll try that, but it will take about a week or two. I've wrote down all the caps type. Anyway, is it ok to replace the 16V 1.200uF electrolytic caps with 16V 470uF polymer caps for the 12V line near the CPU (it has only 2 of these caps)? And is it ok to replace the 6.3V 3.300uF electrolytic caps with 6.3V 1.500uF polymer caps (it has 9 of these caps) .

I *think* that's a Intel-specific thing and not really a problem. Reason I say this is because I have seen many Dells with Intel boards that do this. My Dell Dimension does the same thing when AC cord is disconnected and then reconnected.
I also have a Gateway Select 750 that does this, and its motherboard doesn't have a single electrolytic capacitor (it's all tantalum and ceramic capacitors).

I'm not sure about the other issue, though.

Thanks for the info, i rarely dealing with Intel board lately, i hope that's nothing important as you said .

[HexaPro]

Re: Intel DG965RY power on every time main AC power connected

This is the original caps used for the board:

25V
1x 220uF 8mm Diameter
6x 100uF 6.3mm
1x 33uF 5mm
2x 22uF 5mm

16V
2x 1.200uF 10mm (can i change it with polymer 470uF)
1x 470uF 8mm
4x 4.7uF 4mm

10V
14x 470uF 6.3mm

6.3V
9x 3.300uF 10mm (can i change it with polymer 1.500uF)
2x 1.000uF 8mm
2x 220uF 8mm
4x 220uF 6.3mm KZG (one of them bulging and i have replace them all with 10V 470uF 6.3mm from donor board, can i change it to polymer 470uF?)
2x 100uF 6.3mm

2.5V
8x ???uF 8mmD x 8mmH Polymer, probably Fujitsu (it's the VRM around CPU, i think it's still OK)

Which one i suppose to replace first? If it's possible i want to change it all to polymer. Really need advice for this before i order the caps. Thanks.

[momaka]

Re: Intel DG965RY power on every time main AC power connected

You need to list brand and series for each capacitor type. Otherwise, there's no way of knowing which one need to be replaced and whether polymer caps would work in their positions or not.

The yellow Fujitsu caps are reliable and should be fine. You can leave them alone.

The KZG can be replaced with Rubycon MBZ or Nichicon HM. Rubycon MCZ and Nichicon HN have lower ESR and higher ripple current, so they would likely work fine as well.

[HexaPro]

Re: Intel DG965RY power on every time main AC power connected

Thanks momaka. Ok, here's what i know (the ??? means i'm not sure what brand or type it is).

25V
1x 220uF | 8mm Diameter | Nichicon VR
6x 100uF | 6.3mm | Nichicon VR
1x 33uF | 5mm | Nichicon VR
2x 22uF | 5mm | Nichicon VR

16V
2x 1.200uF | 10mm | Chemi-Con KZE (can i change it with polymer 470uF)
1x 470uF | 8mm | Nichicon VR
4x 4.7uF | 4mm | BP ???

10V
14x 470uF | 6.3mm | Nichicon VR

6.3V
9x 3.300uF | 10mm | Nichicon HM (can i change it with polymer 1.500uF)
2x 1.000uF | 8mm | Nichicon VR
2x 220uF | 8mm | Sanyo NE/NP ???
4x 220uF | 6.3mm | Chemi-Con KZG (one of them bulging and i have replace them all with 10V 470uF 6.3mm Nichicon VR from donor board, can i change it to polymer 470uF?)
2x 100uF | 6.3mm | SME/BP ???

2.5V
8x ???uF | 8mmD x 8mmH | Polymer (red color | 691/561), probably Fujitsu (it's the VRM around CPU, i think it's still OK)

If it's possible i want to poly mod all the caps, need advice about which one and what uF are allowed.

For other caps that can't be poly mod what brand and type is good? I'm planning to use Samxon but i'm not sure which type is the best, since all of them seem to be low ESR so it left me thinking whether to choose the high ripple or the long life? Isn't high ripple = "long life" too?

I have previously recap MSI P31 Neo busted VRM Electrolytic OST with Polymer X-CON ULR and it works fine now. Do you think X-CON is a good reliable brand?

[PCBONEZ]

Re: Intel DG965RY power on every time main AC power connected

Some boards have a jumper to force power-on when power is available.
[This is mostly on server boards.]

Others [probably most] have a setting in the BIOS that does the same thing only it's intended/worded to be for the PC's desired condition after a power failure.
Usually like:
Setting: Condition after power failure
Choices: 'on' / 'off' / 'previous condition'.
.
Yours is probably set to 'on'.
.

[PCBONEZ]

Re: Intel DG965RY power on every time main AC power connected

Unless I missd something the only caps you really need to change are the KZG [all] and any HM with bad date codes. [2001-2004]

16V
2x 1.200uF | 10mm | Chemi-Con KZE (can i change it with polymer 470uF)

2x 470uF is not enough uF there.
Normal on factory poly boards would be 4x330uF or 3x470uF.
If you only have two 'spots' then use at least 2x680uF if not 2x820uF
.
Those you wrote as BP might be bipolar.
If there is no polarity stripe on the side that is why.
.
Are all nine of the 3300uF HM in Vcore?
.

[HexaPro]

Re: Intel DG965RY power on every time main AC power connected

Others [probably most] have a setting in the BIOS that does the same thing only it's intended/worded to be for the PC's desired condition after a power failure.
Usually like:
Setting: Condition after power failure
Choices: 'on' / 'off' / 'previous condition'.
.
Yours is probably set to 'on'.
.

Nope, i know about that and i always set it off.

Unless I missd something the only caps you really need to change are the KZG [all] and any HM with bad date codes. [2001-2004]

The board dated 2007.

2x 470uF is not enough uF there.
Normal on factory poly boards would be 4x330uF or 3x470uF.
If you only have two 'spots' then use at least 2x680uF if not 2x820uF
.

Phew... Thank God. I just ordered 16x Polymer Nichicon LG 16V 820uF from this site. I'm glad i didn't rush and order Polymer X-CON ULR 16V 470uF from capsmod.net that's the highest uF they got for 16V Poly.

Those you wrote as BP might be bipolar.
If there is no polarity stripe on the side that is why.
.

I guess so, i just notice there's no white strip on it or on the board PCB.

Are all nine of the 3300uF HM in Vcore?
.

I'm not sure about it (is Vcore = VRM?), anyway here's the picture.

[PCBONEZ]

Re: Intel DG965RY power on every time main AC power connected

Vcore is the CPU side of VRM [output of VRM].
They will all be in parallel which you can check with a DMM by doing basic continuity checks. The + side goes to the CPU Vcore. The - side goes to ground.
Also these are usually very very low ESR.

VRM input caps [VRM-High or VRM-In] are also all in parallel. [With each other.]
Those will have the + connected to +12v PSU-CPU power connector pin(s).
The - will be on ground.
Almost always 16v or more caps for voltage. **note
Judging by your caps list those would be the 2x1200uF 16v KZE on your board.
Something with ESR like KZE or ZL is not uncommon there but caps with much lower ESR are more typical in recent years.

**note: Very old boards [like P3 and older] usually powered the CPU from +5v so the VRM-In caps voltage rating might be as low as 6.3v.
Intel actually recommended using +12v to power CPUs since the [P3] i810 chipset but almost no one including Intel [except maybe on servers] actually switched to powering CPU's from +12v until P4.
.

[PCBONEZ]

Re: Intel DG965RY power on every time main AC power connected

The red marked Fujitsu Polymers next to CPU socket are the Vcore caps.
The "F" with bars above and below is their brand logo.
Don't need to touch those.

You HM's don't need replaced on a 2007 board unless heat got them and they bloated.
Also outside of the VRM using polymer without checking the circuit for -each- cap can lead to problems. [Like draw out the schematic and do the math.]
Info on the web suggests it usually works, and it might - USUALLY, but people that screw-over their board with a mod usually don't hop on the web and report their stupidity to the whole world as often as people that got lucky and it worked.
[If you don't see a success story for your exact board then don't assume it will work.]

The KZG need to go.

If any of the small 85C caps are bloated -even slightly- they should all go.
85C caps don't live as long as 105C caps [about 1/4 as long if they have the same Endurance Lifetime rating.] and if the box is over about 4-5 years old or if it's a 'hot box' [not well cooled] those can be going bad.
- At least check them closely.
- Small caps don't bloat much when they do bloat. Hard to see the bloat.
Don't let the small size fool you.
While most don't, -some- of those small caps have really important jobs.
.

[HexaPro]

Re: Intel DG965RY power on every time main AC power connected

Judging by your caps list those would be the 2x1200uF 16v KZE on your board.

That's exactly what i'm thinking too. So any 16V near the CPU (between CPU and the I/O back panel) are usually the VCORE. Thanks alot PCBONEZ.

So right now i just have to try replace the VCORE electrolytic Chemi-Con KZE 16V 1.200uF with polymer Nichicon LG 16V 820uF and electrolytic Nichicon 6.3V 3.300uF with polymer X-CON ULR 6.3V 1.500uF.

I'll post the result when it's done and properly tested. The shipment to Bali would take about two weeks or more. I hope the problem with the slow CPU loading has gone by then. If not i'll try to replace all the other caps and hopefully it would be polymer.

I have a hardware service and replacement contract with my customer, so i need to provide the most durable parts that i won't have to replace the whole nice system just because some lousy caps have gone mad.

[PCBONEZ]

Re: Intel DG965RY power on every time main AC power connected

That's exactly what i'm thinking too. So any 16V near the CPU (between CPU and the I/O back panel) are usually the VCORE.

NO. - That's backwards.
The 16v caps are the VRM Input.
Vcore is VRM output.
It's called Vcore because it's the core voltage for the CPU.

[PCBONEZ]

Re: Intel DG965RY power on every time main AC power connected

So right now i just have to try replace the VCORE electrolytic Chemi-Con KZE 16V 1.200uF with polymer Nichicon LG 16V 820uF and electrolytic Nichicon 6.3V 3.300uF with polymer X-CON ULR 6.3V 1.500uF.

Neither the KZE or the 3300uF Nichicon should need replaced unless they are bloated but you can if you want.
-
Those Nichicon are NOT in VRM and reducing the uF is NOT advisable short of doing a complete circuit analysis for each cap to insure the circuit will still work at the new value.

I would NOT replace the 3300uF with 1500uF.
I don't care if the new ones are poly or not.
The *uF* probably MATTERS there.
The result may be a hopelessly unstable board.

VRMs are a known entity. They are all pretty much the same as far as the caps go.
One set of caps is on the PSU and the other set is on the CPU. [PSUs and CPUs need the same no matter what board.]
Plus you can usually look at factory all poly boards to see what works for a given CPU type.

The other circuits on boards are not so similar between different boards.
For anything other than VRM you need to analyze the circuit by doing the math for any changes.
That means drawing out the schematic and looking up the ICs and playing with your calculator.
- THAT's what I tried to tell you before.

Poly modding a lytic VRM is straight forward and what works is known.

Poly modding the rest of a lytic board is either easter-egg hacking and hit-miss dumb luck -or- a HUGE project.
.

[PCBONEZ]

Re: Intel DG965RY power on every time main AC power connected

Have you tested this thing with a different PSU?
The PSU may be crap or have bad caps. [Or both.]
.

[joshnz]

Re: Intel DG965RY power on every time main AC power connected

My dell optiplex does the same. it seems to be a test.

[HexaPro]

Re: Intel DG965RY power on every time main AC power connected

NO. - That's backwards.
The 16v caps are the VRM Input.
Vcore is VRM output.
It's called Vcore because it's the core voltage for the CPU.

Don't really understand what you're trying to explain. So VCORE voltage goes directly to the CPU without passing any more caps? Got any picture of it?

Neither the KZE or the 3300uF Nichicon should need replaced unless they are bloated but you can if you want.

Unless they are bloated or not bloated?

Those Nichicon are NOT in VRM and reducing the uF is NOT advisable short of doing a complete circuit analysis for each cap to insure the circuit will still work at the new value.

I would NOT replace the 3300uF with 1500uF.
I don't care if the new ones are poly or not.
The *uF* probably MATTERS there.
The result may be a hopelessly unstable board.

Ok, but if i insist to try would that damage the board or other still good caps?

VRMs are a known entity. They are all pretty much the same as far as the caps go.
One set of caps is on the PSU and the other set is on the CPU. [PSUs and CPUs need the same no matter what board.]

Got any picture? I'm new at this and i don't really know what you are saying.

The other circuits on boards are not so similar between different boards.
For anything other than VRM you need to analyze the circuit by doing the math for any changes.
That means drawing out the schematic and looking up the ICs and playing with your calculator.
- THAT's what I tried to tell you before.

I'm not good at math T_T!

Poly modding a lytic VRM is straight forward and what works is known.

Yup i think so too.

Poly modding the rest of a lytic board is either easter-egg hacking and hit-miss dumb luck -or- a HUGE project.
.

Yup, but how long does a good lytic caps would last for 14/7 operation? I'm obsessed with poly since it seem very durable. I always suggest my customer to buy an all solid cap board so i don't have to worry about complain.

Have you tested this thing with a different PSU?
The PSU may be crap or have bad caps. [Or both.]
.

Yup, didn't i mention before that everything is fine after i replace the board with MSI P31 Neo? I didn't replace anything but the board. Same CPU, RAM, GPU, PSU, HDD, etc. In fact i'm using it right now to post in this forum. Seem ok, but i have never try it for my customer Fuji software again. At least until i replaced all the 16V and 6.3V caps. I don't want to interrupt his business without knowing that it's worth a try.

My dell optiplex does the same. it seems to be a test.

Hm... Ok, now i think that is Intel's unique feature

[PCBONEZ]

Re: Intel DG965RY power on every time main AC power connected

This pic [or ones that show the same thing] is already in the forum about 30 times over.
-
It's WAY over simplified but it shows what you are asking.
[Drawing is missing components and there it's only one Phase. Most VRM's since PIII have 3 or more Phases.]

Technically the Vcore caps are in parallel with somewhere between 10 and 30 very small [small both in size and uF] caps located in the CPU or in the CPU socket. Exactly what and where depends on CPU type. Those are there to handle high frequency EMI [like from a radio, a microwave or a cell phone] and they don't do enough at Ripple Frequency to be concerned with their effects.
.

[HexaPro]

Re: Intel DG965RY power on every time main AC power connected

Hi PCBONEZ, my caps have arrived.

I've replace the following caps :
2x 16V 1.200uF D10mm UCC KZG with Nichicon LG 16V 820uF,
4x 6,3V 220uF D6.3mm UCC KZG with Samxon X-CON ULR 6,3V 470uF, and
9x 3.300uF 10mm Nichicon HM with Sanyo OS-CON SEPC 1.500uF.
Everything seem run faster and fine so far.

But today i've seen that a chip has been burnt, i don't know what chip it is? Does this has any bad impact for the system? I've attached the picture too.

I want to proceed the poly mod with the
14x 10V 470uF 6.3mm Nichicon VR With Samxon X-CON ULR 6,3V 470uF
but i doubt that the board worth it, unless i'm sure that the damaged chip won't give any problem.

The system sometimes would BSOD memory dump or unknown hard error before the recap. I don't think that my RAMs are faulty, they were working fine before (Is it because that burned chip?) Should i proceed and replace all the 10V D6,3mm, they are gather around the DIMM mostly. Would a faulty caps will cause BSOD memory dump? The unknown hard error probably caused by the hard drive (i think it's faulty).