Dell Precision 470 Workstation

[69ragtop]

Re: Dell Precision 470 Workstation

So I just came across an old 470, caps were all blown to crap. I ordered a cap set from the CapKing and went to work on it yesterday.

Originally I got an error about the VRM being faulty. If I pulled the second CPU and it's VRM, I got an error about the other VRM being faulty but the system would power up (at least 12v side) and the fans would go crazy.

After re-capping, nothing, I get a solid B light showing a power supply issue. I know that isn't the case as the power supply already fired up, unless the 3.3v or 5v are dead.

I did have two issues putting the new caps in, two of them just wouldn't go through the holes smoothly so I left the iron on them a bit too long I think, they were hot but not HOT, however I did smell a smell I've never smelt before, acrid, I think I may have burned one of the caps.

Any thoughts on what I can try? I don't want to drop any more cash into this so I'm probably going to trash it unless someone has any ideas. Thanks!

Did you ever resolve the issue? I have a similar machine. With one PS, I get green/amber/amber/off which indicates VRM0. It is a single processor. When trying another PS I see off/amber/off/off (i.e. solid B light). However, both power supplies will power a different 470 with no issues.

[michaelahess]

Re: Dell Precision 470 Workstation

So I just came across an old 470, caps were all blown to crap. I ordered a cap set from the CapKing and went to work on it yesterday.

Originally I got an error about the VRM being faulty. If I pulled the second CPU and it's VRM, I got an error about the other VRM being faulty but the system would power up (at least 12v side) and the fans would go crazy.

After re-capping, nothing, I get a solid B light showing a power supply issue. I know that isn't the case as the power supply already fired up, unless the 3.3v or 5v are dead.

I did have two issues putting the new caps in, two of them just wouldn't go through the holes smoothly so I left the iron on them a bit too long I think, they were hot but not HOT, however I did smell a smell I've never smelt before, acrid, I think I may have burned one of the caps.

Any thoughts on what I can try? I don't want to drop any more cash into this so I'm probably going to trash it unless someone has any ideas. Thanks!

[bigcape]

Re: Dell Precision 470 Workstation

Thermistors are used as inrush-current limiting devices in PSU.
You probably had a good strong power surge at some point.
It may have thrown setpoints off or toasted other parts like the controller chips.
Certainly the PSU could have had bad voltages and that fried your caps.
.

The system was hooked up to a ups so it had to be PSU side then. They sure did put a lot of white goop in thi thing. That stuff alone looks like bad air flow in itself.

Here is a pic

[PCBONEZ]

Re: Dell Precision 470 Workstation

I see a Thermister NTC1 is cracked. Looks like it let a little "bang-poof" onto the white plastic shielded card. The rest looks fine.

can just that one varible resistor be enough?

Im going to take off that big peice of copper to take a peek at the mack-daddy cap of them all! first im going to find my bleeder resistor

Thermistors are used as inrush-current limiting devices in PSU.
You probably had a good strong power surge at some point.
It may have thrown setpoints off or toasted other parts like the controller chips.
Certainly the PSU could have had bad voltages and that fried your caps.
.

[bigcape]

Re: Dell Precision 470 Workstation

You have three +12v rails numbered A, B and C.

Ah thank you! Did you see my post about the thyristor above?

[PCBONEZ]

Re: Dell Precision 470 Workstation

PCBONEZ,

Whats this 12VA, 12VB, 12VC stuff? I understand the others

You have three +12v rails numbered A, B and C.

[PCBONEZ]

Re: Dell Precision 470 Workstation

Could you post that or provide a link? Enjoy a good read.

Here's the thread:
https://www.badcaps.net/forum/showth...&highlight=mod

Here's the document:


See page 15.

[bigcape]

Re: Dell Precision 470 Workstation

PCBONEZ,

Whats this 12VA, 12VB, 12VC stuff? I understand the others

[PCBONEZ]

Re: Dell Precision 470 Workstation

Could you post that or provide a link? Enjoy a good read.

KC8 posted it like 3? 4? years ago.
Always takes me forever to find it but I'll look for a few before I crash for the night.

[bigcape]

Re: Dell Precision 470 Workstation

First crack it open .

I see a Thermister NTC1 is cracked. Looks like it let a little "bang-poof" onto the white plastic shielded card. The rest looks fine.

can just that one varible resistor be enough?

Im going to take off that big peice of copper to take a peek at the mack-daddy cap of them all! first im going to find my bleeder resistor

[PCBONEZ]

Re: Dell Precision 470 Workstation

So where is the lytics to poly conversion chart?

There isn't one because too much is involved.

But you are welcomed to do the math [for each cap] if you want to.
Here's a start.
.


.
I'm of course just teasing.
The point is that it isn't that simple.

.
There is a thumb-rule used for VRM caps derived from that math and an engineering document for a VRM MOSFET Controller chip that says if you switch to ploy then you can reduce uF by half. That doc clearly pointed out that in VRM's the ESR is more important than the uF.
-
The thumb-rule isn't that safe to use anymore because it was based on the characteristic of the polys that existed at the time. Since it was created more kinds of polymer have come on the market and some actually have worse ESR than the better lytics.

.

[Toasty]

Re: Dell Precision 470 Workstation

...
- The rule about reducing uF when you lower ESR does not apply to other parts of a motherboard.
[That rule wasn't 'invented' out of the blue, it comes from a tech doc about VRM's.]...

Could you post that or provide a link? Enjoy a good read.

[bigcape]

Re: Dell Precision 470 Workstation

First crack it open [*with it unplugged] and look for blown caps and/or charred darkened components or PCB.
* With it energized the heat-sinks can be at 2x line voltage or more.

I'm GOING in!

A power supply with a handle ought to tell one something all by itself. Especialy in a workstation series!

[PCBONEZ]

Re: Dell Precision 470 Workstation

(bold- bigcape)

Dell's pinouts on their PSU's I've been told are a standard of their own. How would you test the old PSU. I've got it setting right here!

Thx

Sean-Bigcape

First crack it open [*with it unplugged] and look for blown caps and/or charred darkened components or PCB.
* With it energized the heat-sinks can be at 2x line voltage or more.

[bigcape]

Re: Dell Precision 470 Workstation

I'd give that board a FULL poly-mod. If it gets re-capped in lytics, won't it just happen again, since the busted caps are rubycon?
I think that re-capping a board with busted rubycons in lytics is like replacing bad Fuhjyyu caps with more Fuhjyyu. It would fix it at first, but you would have to recap it again in a year or so.

So where is the lytics to poly conversion chart?

[bigcape]

Re: Dell Precision 470 Workstation

If you haven't checked the PSU then you may.. (have).a PSU that blows caps.

(bold- bigcape)

Dell's pinouts on their PSU's I've been told are a standard of their own. How would you test the old PSU. I've got it setting right here!

Thx

Sean-Bigcape

[bigcape]

Re: Dell Precision 470 Workstation

And, as has been said several times now...
Odds are those caps blew because you had bad cooling or a bad PSU.
If you haven't checked the PSU then you may be running your new board on a PSU that blows caps.

The 470 does not auto shut off if it was caused by bad cooling?

I actualy started by changing the PSU before the blown caps were discovered.

It will be interesting to crack open the OLD PSU!

How do i assure that is not any stored power in it? Are there bleed points where I can short out with a resistor or is it a self beeding PSU?

I never had "operated" on a PSU, just "test n replace"

[PCBONEZ]

Re: Dell Precision 470 Workstation

That board isn't known to frequently blow caps outside of the VRM Vcore.
- VRM Vcore is where the polys are in the pic I uploaded.
- That's why Dell changed to polys there.

And, as has been said several times now...
Odds are those caps blew because you had bad cooling or a bad PSU.
If you haven't checked the PSU then you may be running your new board on a PSU that blows caps.

VRM Poly upgrades are easy.

"Full Poly Mods" aren't that simple.
A poorly done full poly mod can lead to an unstable system or even damage the board.
- The rule about reducing uF when you lower ESR does not apply to other parts of a motherboard.
[That rule wasn't 'invented' out of the blue, it comes from a tech doc about VRM's.]
- Some x-mm at x-uF at x-many volts combinations are expensive, hard to find or simply non existent. Poly's max out at about 1800uF/6.3v and 1000/16v. To get higher uF caps you have to go down in volts.
- And you can't cheat on [lower] the volts in other areas of the boards like you can in VRM Vcore because, without physically checking, you don't know the actual circuit volts. You will have to have a working board and actually measure the voltage on each cap with the system loaded and running in each operating condition before you know what voltage replacement caps you can get away with in each position. [Yes, check for every cap.]
-

[bigcape]

Re: Dell Precision 470 Workstation

If my board is already a poly mod or a later production (im going for the latter because the welds are not human!), the other caps's are the reveled Rubycon's...... Then

How do you test if here is a possibility that the VRM chip is damaged because a FET has shorted and therefore Vcore will be too high?

This type of testing and how is new and unexplored by me!

[c_hegge]

Re: Dell Precision 470 Workstation

I'd give that board a FULL poly-mod. If it gets re-capped in lytics, won't it just happen again, since the busted caps are rubycon?
I think that re-capping a board with busted rubycons in lytics is like replacing bad Fuhjyyu caps with more Fuhjyyu. It would fix it at first, but you would have to recap it again in a year or so.