Recapping an Albatron PX865PE Pro II for Albuquerque

[Super Nade]

Albuquerque sent me a pm asking me to take a look at his board and determine if it warrants a recap.

I'd leave it up to him to accurately describe the symptoms and issues he was having with the board. Without further ado, let us take a look at this board.

By visual inspection, I can say that none of the caps are leaking or bulging at this point.

The caps are as follows:

• CPU VRM side: Sanyo WG (K-Vent) and OS-CON (Purple)
• Chipset: GSC and Panasonic (T-Vent)
• DRAM: Panasonic (T-Vent) and GSC

All the crap GSC's are going to be replaced and the empty spots near the inductor are going to be populated.
Analysis will follow in an upcoming article. I'll be including O-scope signal screens and if I can finish building the probe, you may see some spectrum analyzer shots as well.


In the meantime, here are a few pictures.

Slideshow:

http://img504.imageshack.us/slideshow/player.php?id=img504/6983/1197660536q3n.smil

-------------------------------------------------------------------------------------------------------

 

[Super Nade]

Phase-I: The initial test

Something is definitely wrong with the board. Memtest errors out with a register dump loop. I tried different floppies, different FDD's and different cables. I also cleared the CMOS, beefed up the voltages a little bit, tried different slots on the MoBO, tried a single stick of RAM, swapped them around and also tried a generic crap Value RAM stick I had laying about. No go.

I'm gonna try a CDROM drive, but unless I see memtest running, at this point, I'm calling the board suspect and will proceed with the next phase.

 

[Albuquerque]

That's odd... Haven't had memtest errors before. I even tested probably a month ago before I installed Vista. Even after resetting the BIOS options back to stock timings and whatnot?

The original reason for sending this out was two things:

First thing: The S478 Presc"Hot" processor in there is good for 4.2Ghz under my water cooling, but the overclock has been slowly dwindling over the last six to eight months. At first I assumed it was the processor degrading, with the 1.575v and all. It's most recent stable OC has been a skimpy 3.6Ghz. However, I have another of these exact boards that has almost zero use, and the CPU overclocked right back to 4.2Ghz as soon as I swapped boards.

I know what you're thinking, why not keep the other board? Because the "new" board was an earlier revision that had problems with the AGP lock after about 255FSB. So to get 280FSB, I had to run the AGP and PCI busses at a whopping 93/46. I ended up giving that board to my brother (who desperately needed a PC to replace his P2/333, 256mb Compaq system) with my old 3.0C Northwood that worked great at 3.6Ghz. It was perfect for a 240FSB setup, and since the board was nearly new, it was the better choice to give to him.

Ok, so back to my board: Second reason to send it in? Because it's a fantastic board. I don't know why people never paid attention to Albatron, especially their PX845PE Pro IIs or this PX865PE Pro II. This board ran a 2.4C at 292FSB with memory at 1:1 (yes, DDR584 memtest stable for days) three years ago. All at 1.6VCPU, 2.85VMEM and 1.8VAGP. Dual BIOSes, dual RAID implementations (one for SATA, one for PATA), firewire, USB, gigabit LAN on Intels' CSA bus (250mbps connection to northbridge for full gigabit transfer capacity, unlike most gigabit adapters which are connected to the 133mbps PCI interface). ALso came with Envy24 7.1 audio -- the same basic chipset used in the M-Audio Revolution and the like.

It's an excellent, no, phenominal motherboard in my opinion, and there simply isn't anything out today that has the same amount of quality features no matter the pricetag. And since this puppy (when running at full speed) does everything I want at speeds that are quite reasonable to me, I don't have any reason to upgrade yet.

So, I want this board to live another 12-18 months so I can see how Nehalem performs (and whatever AMD sends as competition) and finally upgrade. But I just love this board, so I want it to live a while longer before it goes to another home, likely my mom

 

[Albuquerque]

Oh, and that 2nd to last picture? Look at the three caps that "surround" the AGP slot -- one in front, one in the middle on the opposite site near the northbridge fan, and one at the back on the same side as the first... All three have the little Y-shaped vent in the top.

Those are the caps that I assumed were bad, because they all appear to bulge to me. And not all of them appear to bulge the same amount, if that makes sense. Further, there's another of those caps (in that same pic) between the 1st and 2nd PCI connector that doesn't look bulged like the rest, but looks to be same manufacture and design.

So, are there simply some caps that are built like that? How do you tell the difference?

 

[Super Nade]

They are not bulging. I went over them with a magnifying glass. A bulging cap would look like this-->http://super-nade.blogspot.com/2007/11/recapping-dead-epox-8kra2.html

I'm stuck at home today because of the bloody snow. It's been gusting since 4 AM EST.

 

[corruption]

Super Nade: I love the detail that you got into in this post (Here) without going completely overboard with various calculations. (I'm a Computer Electronics Technician and can appreciate how long that must have taken to write and keep it relatively simple.) If I ever need help choosing replacement capacitors for a motherboard I know who to ask!

 

[Super Nade]

Thank you for the kind words. It took me several drafts (I'm still not happy with the published version) because there was just too much information. Should I talk about ESL? should I bring in n-phase VRM topology? Should I talk about transients? Should I discuss consequences of adjusting the feedback loop parameters of the DC-DC converter? All these are pertinent, but then again, if I begin to indulge myself, I'll probably never get the job done.

As a compromise, I put up the table for quick reference because the (culled) write-up was still a bit too much of an information overload. 99% of our members who read the front page give what I write the miss because it is a bit too detailed. But unfortunately, I can't bring myself to skip details or not cover all bases.

Curiously, I receive a lot more feedback from members on other message boards.

 

[Albuquerque]

Cool, well thanks for the lesson on bulge. I know I had problems with memtest stability over a certain speed, but it's still odd that it fails straight away. I don't believe I was using the newest version of Memtest; I want to say it was version 3.2? Could that be a valid reason why I never saw any obvious problems versus yours that won't even start? Not sure what the difference would be though :s

Well anyway, I'll keep watching and replying. Thanks a ton for looking at it!

 

[Super Nade]

Well, I was going to try the old memtest today, but for the snow.

 

[Super Nade]

Phase-II

No go man. Old memtest is erroring our as well.

Here is a picture slideshow of a few basic O-Scope shots of the DRAM FET signal. Yeah, everybody knows that the RAM runs off of the 3.3V line, but here is some proof that the signals do not exceed 3.3V.

I'm going to post detailed analysis of everything I did so far on my Blog (see sig), only because I started posting there and it is well arranged in a readable format.

So, far apart from everything getting a bit warm, nothing really crazy is going on. Ripple is a bit high, but that can be due to the ghetto and wholly inadequate probes I'm using.

This freaking imageshack upload is bloody slow, so I'll update this post in an hour or so.

Updates:
http://img127.imageshack.us/slideshow/player.php?id=img127/731/1197837852ok3.smil
http://img134.imageshack.us/slideshow/player.php?id=img134/7729/1197837203te2.smil
http://img127.imageshack.us/slideshow/player.php?id=img127/731/1197837852ok3.smil

 

[Super Nade]

Final update for the day

1. I found some very interesting things that should be kept in mind when replacing caps.
2. OS-CONs help with voltage stability, no doubt. There is a large difference between using OS-CONs v/s something else.
3. Just as I said in my earlier articles, it is sufficient to use 2.5V rated caps on the CPU transient suppressing side (the bulk cans). Ripple current rating/handling-capacity is more important than raw capacitance value alone.

 

[Albuquerque]

I'm seriously wondering if it was somehow damaged in shipment...

I say this, because that board was taken from a fully working system. I mean, it wasn't operating at the speeds I would like, but I was playing Crysis for multiple hours at 3.6Ghz CPU / 400Mhz DDR the same day that I took it out to ship to you.

I have to assume that if it was this unstable, I would have never made it into the operating system. Now I'm a bit worried that somehow DHL dropped it, kicked it, submitted it to about 8KW of EMI somehow, et al.

Oh, one more thing: the CPU and heatsink are connected by whatever the leftover heatsink compound was left from the waterblock. I entirely assumed you'd be instantly removing the heatsink for a better look at the VRM caps... So with that in mind, how do the temp readouts in BIOS look?

 

[Super Nade]

I didn't turn it on long enough to get accurate temperature readings. However, I'm gonna try an old skool memtest as a final attempt. The box arrived in good condition, so shipping damage was unlikely. I somehow think it has something to do with the version of memtest I'm using. I will keep trying.

Dude, I don't know how you can run this setup without having at least 3 other fans on the MB. I would sink all FETS, get rid of the crap stock HSF and drop an 80mm on the RAM.

 

[jmh547]

Albuquerque I just realized what your sig says...

 

[Albuquerque]

I didn't turn it on long enough to get accurate temperature readings. However, I'm gonna try an old skool memtest as a final attempt. The box arrived in good condition, so shipping damage was unlikely. I somehow think it has something to do with the version of memtest I'm using. I will keep trying.

Dude, I don't know how you can run this setup without having at least 3 other fans on the MB. I would sink all FETS, get rid of the crap stock HSF and drop an 80mm on the RAM.

Heh, actually it's been quite nice just the way it is for a considerable amount of time. Of course, the CPU is watercooled which helps tremendously with the area surrounding the CPU. My (admittedly short) testing of the stock cooler netted something around 65*c under full load at 3Ghz. I typically see no more than 52*c under full load at 4.2Ghz. Also, the 120.2-esque radiator is in the front of the case, so the dual 120mm (fanbus-controlled) fans pulling through the rad blow straight across the board.

And the memory? I didn't think they were too hot... When I had my Komusa HyperRam at DDR584 @ 2.85v there was some pretty intense heat, but cooling them made no difference to the clocking. For a short while I also had some Crucial Ballistix in there, and it ran only a bit cooler than the Komusa. I've had the Geil sticks for about six months, and they've run notably cooler than either of the above, especially at the (below-spec) 2.55v I run them at.

Oh, and I assume you're talking about the northbridge HSF? Yeah, about two months ago the sleeve bearing started to go out. I've taken it out and re-packed / greased the bearing and it seems to be doing better now. In reality, the northbridge temps have never been an issue -- they hover anywhere between 35-45*c for me.

 

[Super Nade]

Oh the horror!

I was finishing up the recap and so I powered on the board to check for a voltage and then I saw a flash! Shocked beyond belief, I yanked the power chord as quickly as I could move. WTF just happened??? All voltages were within spec?

It so turned out that a piece of leftover metal after cutting the legs of a cap was stuck in my hair and it fell on the board as I bent down to take a closer look.

I mean how unlucky can one get? I think I owe Alberquerque a new motherboard. Sorry mate.

Upon inspection, I saw that the ground trace blew. So I fixed it up, but I have not tested it yet. I'm waiting on a shipment of Chemicon LXZ's before I can test that.

 

[Albuquerque]

My first reaction was LOL. As in, what else can you do but laugh?

Not really sure anyone even sells these boards anymore; a quick glance around the 'net suggests they really don't. I think the ONE place that has them wants way more than they're worth. I can only imagine what my Vista license will think of a new board

So... What was that trace grounding? I mean, other than the obvious damage to the trace itself, what else was downstream? If all you ended up doing was popping a FET or another cap, then I'm pretty "meh" about it.

I'm not the panicky type; give 'er another go when you get the appropriate parts in. Just make sure to comb the hair next time

 

[Super Nade]

No FETS are damaged. I may have popped a cap, but I'll replace all the caps just to be sure. As far as I can tell, the only thing that blew was the trace. This is because of the short caused by that stupid wire. Next time, I'll wear a hat and make sure I use compressed air just to make sure.

At this point, I'm even ready to do a few push-ups before I begin. :

 

[Albuquerque]

Seriously man, I'm not worried. I'm sure there are people out there (even on this forum) who'd yell, scream and call for your azz... That's not me Continue cookin' away, and it should all turn out fine.

And keep postin' pics

 

[Super Nade]

Here is the plan ; I'm waiting for a few odd value caps from digikey. Once I have them, I'll go ahead and finish the job, repair the trace (not the hack job that the pic shows) and test the board. Once I see the board posting, I'm going to figure out how to secure the broken trace. Glue or caulk comes to mind, but I'm going to see if I can borrow something else from the Material Science lab.

In any case, I'm going to compensate you in some way or the other for this goof up.

The next update will be next week and I'll be able to post any replies to questions only after this Friday, because I have an urgent family matter to attend to.