Bad Capacitors A7N8X Deluxe

[Tech_Cowboy]

I had my A7N8X with an Athlon 2500+ overclocked for about two years with no problems. Then recently while I was encoding a video, it just shut down. I tried restarting it, but it would go into an endless reboot loop. Nothing I did would bring it back. I even let it cool off overnight. The only way I could get it to boot up properly is to set the BIOS back to its factory timings. Once I did that, it ran just fine. And it still does. But if I try to overclock, it just keeps rebooting itself.

This got me thinking of something I noticed just over a year ago during one of my regular cleanings. Here's a picture of it:



I was a little concerned when I first noticed this, but the machine ran fine overclocked, so I hadn't really thought about it until now. Prior to the unexpected shutdown, there were no reboots, no BSoD, or any other problems. But now I can't overclock, so I'm thinking the caps might be the source of the problem. Does anyone else think so? I thought that if the caps were bad, it wouldn't even run at factory speed. But this machine does. In fact, I'm using it now to post this message. So I don't really know what to think.

Anyhow, I do want to replace the caps. Does anybody know which ones I should replace? All of them? Just the ones that are leaking? Just the larger ones? I checked over at www.badcaps.net but they don't sell a kit for the A7N8X. I'd really like to have a list of the quantity, sizes, voltages, etc that I will need before I start unsoldering the existing ones.

Thanks,

TC

[mydian]

I would replace all of them! If it comes down to it look at the values directly on the capicitor and order them. You may have to measure them to make sure you get the correct size.

For work I order all my components from http://www.mouser.com

[Gibby]

...But now I can't overclock, so I'm thinking the caps might be the source of the problem. Does anyone else think so?

Yes - your caps are obviously are - or were - cooking off their electrolyte through the top of the capacitor as evidenced by the brownish patches. That is either dried electrolyte or corrosion from leaking electrolyte. This may or may not be related to the defective cap issue (more at the bottom of this post).

I thought that if the caps were bad, it wouldn't even run at factory speed. But this machine does. In fact, I'm using it now to post this message. So I don't really know what to think.

It depends. Capacitors can fail in an immediate and absolute way, e.g., when they expode. But some capacitors can also fail progressively over time - minutes, hours, days, months - drifting further and further out of spec. Those will either completely die at some point, or else eventually drift so far out of spec that the circuit as a whole 'fails'. There are also several interrelated 'specs' of a capacitor. Which one/how far from spec = failure is another 'it depends' question. A failing or completely failed capacitor *might* never leak. However, a leaking capacitor generally assures an immenent failure.

I think you're seeing the latter case: some of your CPU power caps are almost fried. Good enough for stock clocking, but not good enough to handle the power quality needed for overclocking.

Anyhow, I do want to replace the caps. Does anybody know which ones I should replace? All of them? Just the ones that are leaking? Just the larger ones?

I would start off with 'just the larger ones' - they're used by the crappy two-phase A7N8X CPU VRM power circuit. The smaller caps elsewhere on the board are much less likely to be an issue unless they are clearly leaking. There are eight total of the 'big' capacitors: your diagram shows a row of five and a row of three; you have bad caps (your red arrows) on both 'rows'.

The 'big' caps that are not visibly leaking may still be bad, so I would go ahead and replace them. There is no simple way to test them while they are still soldered to the motherboard, and you probably don't have the proper equipement anyway. You cannot simply test capacitance - it's a little more complicated than that.

ASUS switched to the better Rubycon caps at some point. [Rubycons look like they have a 'K' rather than an 'X' scribed on their safety blow-off silver top. Not definative for ID, the 'X' just means an A7N8X absolutely does NOT have Rubycons, a 'K' *might* but does not positively indicate Rubycons] You can start off trying to find the Rubycons for replacing your current capacitors - or use their specs for determining an equivalent brand/model.

I checked over at www.badcaps.net but they don't sell a kit for the A7N8X.

Use their form to order a custom kit, then just specify eight caps with the specs below. They don't need every spec listed - just give them the ones they will ask for via the dropdown menus.

Rubycon MBZ-series 3300uf 6.3VDC. Rubycon's MBZ-series have high temp ratings (105C), ultra-low z [= ultra-low ESR or impedence] and can handle high ripple current. They are rated at 2000 hours at maximum continuous load - don't freak out about that being too low. The CPU voltage regulation module (VRM) circuits shoudl not be filtering the maximum possible ripple load in and out at 105C for a week straight.

The MBZ-series was made specifically for _M_other_B_oard low_Z_ (impedence) use. Maybe someone else with more electronics knowledge could say if there is either 1) a Rubycon line, or 2) another manufacturer's line with better specs that might make it worth spending some extra money for a 'better' equivalent capacitor. And forget Radio Shack - this is not a general-purpose 3300uf electrolytic capacitor you would find there.

Note that the capacitor leads are polarized (+/-). The vertical white bar running along the side of the capacitor shows the polarity of the lead underneath it. Make sure you orient the replacements appropriately. Ignore the direction of the 'K' on the top - it's orientation is random relative to the polarity of the leads. On the A7N8X-Deluxe Rev. 2 board I have, all the caps in each 'row' are oriented identically, but the row of three caps are all oriented 90 degrees clockwise relative to the caps in the row of five.

Last, but not least, the A7N8X CPU power comes from the 5V rail. Excessively low or high voltage on that rail - or - large amounts of ripple current on that rail stress the eight capacitors in question. It's possible that the caps were perfectly good, but died as a result of a bad 5V rail, i.e., the PSU is to blame. [Edit: corrected original statement that A7N8X gets CPU power from the 12V rail - it absolutely does not, and was actually among the last Socket A boards to continue using the 5V rail for CPU power]

The infamous defective motherboard capacitor issue mighthave solely/partially been to blame, but it's hard to say. I'm not suggesting this to confuse the issue any more - it's just something else to keep in mind. This does not change the fact that your caps should probably be replaced.[/quote]

[Tech_Cowboy]

Rubycon MBZ-series 3300uf 6.3VDC

Great reading. Thanks so much. I'll start with eight of these caps and see if I can overclock again. If that doesn't do the trick, then maybe I'll replace the rest of them. I already knew that the capacitor leads were polarized, but it doesn't hurt to remind anyone else who might be reading your excellent post.

Of course, I will be updating this thread when I've completed the task.

Thanks again,

TC

[Gibby]

I was trying to figure out what brand of caps ASUS used on your board originally, TC. They look like a brand called OST. Do you see that logo on the caps anywhere? It's hard to tell from the angle of the picture.

And speaking of your picture, you might want to think about another mod to your board if you're going to the trouble of replacing those caps. I'm looking at the dust in your CPU fan as well as that on the side of your power caps. Your fan is sucking air from above the CPU heat sink and pushing it through the fins. You're taking what is probably the hottest air in your case and exhausting it right over your power caps, VRM circuitry and northbridge. The back side of your video card acts like an air dam, forcing the hot air up and then probably back over to the CPU fan to be sucked in reheated again. The air tends to just swirl around that area and acts like a nice little convection oven, picking up heat on every pass through the CPU or northbridge fins.

This is especially problematic on this motherboard, since the two-phase VRM components are already stressed. It is more likely to result in unstable power than actually smoke the caps, but in either case it's not good.

I won't bring up the old suck vs. blow debate, but I'll point out that the local airflow around the CPU with stock cooling for this model of motherboard is less than ideal, especially for overclocking. The two-phase VRM power is hard enough on the components and cooling is critical. Ideally - for this motherboard in particular - you want fresh (ambient) air for the VRM components. That's almost impossible without exhausting the CPU and northbridge 'up' and then moving that outside the case as fast as possible.

Way back when the OC crowd was interested in this board, people started noticing how hot the northbridge got at times. The first thoughts were to actively cool the northbridge with one of those little fans you could mount right on top. Which was a great idea, except that most pulled air from the top and pushed it out through the fins.

If you look at the orientation of the northbridge fins, you'll see that they're perfectly placed to exhaust air across your power caps. But on the opposite side that the CPU exhaust was heating the caps. It often turned into a regular cap Bar-B-Que! Guess which two caps usually cooked off first. Even if the caps didn't cook off, the CPU power often became less stable over time as a result. If you look at really old posts on this forum, you'll see instances where people had stability issues that were (at least temporarily) solved by removing the case cover, even though they had plenty of case fans.

The bad cap problems at the time confused the issue more. Most people figured they were just unlucky if the power caps started leaking, "...I must have got one of the bad motherboards..." In reality, I wonder how much of the problems just resulted from questionable design and manufacturers ignoring the airflow issues of a mounted motherboard. No doubt there was a problem with bad caps at the time, but I've seen Rubycons cook off on this motherboard, too.

The motherboard ('case') temp sensor is way over on the opposite corner of the motherboard and does little more than read the ambient air temp entering from the front vents of most cases. If that sensor had been mounted anywhere near the VRM circuitry, you would probably see significantly higher (like 20+C) case temps then you see now. Of course, this would really just be reading your VRM convection oven's temps in the little tornado near the CPU. I would rather have seen that temp though, then getting a somewhat useless reading of relatively fresh air entering the case.

This is probably just all academic now, but it's still an interesting lesson in cooling and case airflow - at least for primates, anyway. I'm afraid I can't speak for species with opposable thumbs.

[Tech_Cowboy]

Gibby,

I don't see OST on these caps. In fact it's hard to read anything on them without pulling the caps or pulling the motherboard out of the case. I'm not quite ready to do that yet. But from what I can see, it looks like they have the letters H-M-M, with the last "M" enclosed in parenthesis, like this:

HM(M)

Does that make sense?

Anyhow, I understand what you're saying about the temperature issue. I thought I was doing a decent job of keeping the temp down inside my case, but after reading what you wrote, I'm not so sure any more. You mentioned a mod that involves exhausting the CPU and Northbridge up and out of the case. Does this mean reversing the fan and cutting a hole in the case? That's not something I'm eager to do, but I'll keep the idea on the table. In the meantime, what about one of those fans that go in an empty PCI slot to pull in more fresh air? Would that be a good short-term solution?

I guess it's not enough to just vacuum out the dust every month or so. By the way, the reason you see dust in that first image because I took that picture right when I opened the case and noticed the leaking caps, which was before running the mini-vac. (And yes, I do remove the CPU fan from the heatsink and vacuum it real well too.)

Speaking of that picture, it was taken almost exactly one year ago. When I looked at the motherboard just now to find the brand name on the caps, I discovered more surprises:



Three more are leaking. And the only one of the "big eight" that's not leaking is bulging (yellow arrow). I've probably looked at these caps a dozen times in the past year and not noticed the new leaks. But when I compare that old picture with the current one, it's obvious that things are steadily getting worse. This might explain why I was able to overclock last year, but not now. As you said, they're probably slowly going out-of-spec. If you're right about the temperature inside the case, then I guess I'm cooking them to death.

TC

[cartel]

Shyt man you seen that a year ago and kept going?
That's like seeing your oil filter in the rear view and driving for a year.
Your lucky as hell you didn't incinerate your components.

[cartel]

I was trying to figure out what brand of caps ASUS used on your board originally, TC. They look like a brand called OST. Do you see that logo on the caps anywhere? It's hard to tell from the angle of the picture.

And speaking of your picture, you might want to think about another mod to your board if you're going to the trouble of replacing those caps. I'm looking at the dust in your CPU fan as well as that on the side of your power caps. Your fan is sucking air from above the CPU heat sink and pushing it through the fins. You're taking what is probably the hottest air in your case and exhausting it right over your power caps, VRM circuitry and northbridge. The back side of your video card acts like an air dam, forcing the hot air up and then probably back over to the CPU fan to be sucked in reheated again. The air tends to just swirl around that area and acts like a nice little convection oven, picking up heat on every pass through the CPU or northbridge fins.

This is especially problematic on this motherboard, since the two-phase VRM components are already stressed. It is more likely to result in unstable power than actually smoke the caps, but in either case it's not good.

I won't bring up the old suck vs. blow debate, but I'll point out that the local airflow around the CPU with stock cooling for this model of motherboard is less than ideal, especially for overclocking. The two-phase VRM power is hard enough on the components and cooling is critical. Ideally - for this motherboard in particular - you want fresh (ambient) air for the VRM components. That's almost impossible without exhausting the CPU and northbridge 'up' and then moving that outside the case as fast as possible.

Way back when the OC crowd was interested in this board, people started noticing how hot the northbridge got at times. The first thoughts were to actively cool the northbridge with one of those little fans you could mount right on top. Which was a great idea, except that most pulled air from the top and pushed it out through the fins.

If you look at the orientation of the northbridge fins, you'll see that they're perfectly placed to exhaust air across your power caps. But on the opposite side that the CPU exhaust was heating the caps. It often turned into a regular cap Bar-B-Que! Guess which two caps usually cooked off first. Even if the caps didn't cook off, the CPU power often became less stable over time as a result. If you look at really old posts on this forum, you'll see instances where people had stability issues that were (at least temporarily) solved by removing the case cover, even though they had plenty of case fans.

The bad cap problems at the time confused the issue more. Most people figured they were just unlucky if the power caps started leaking, "...I must have got one of the bad motherboards..." In reality, I wonder how much of the problems just resulted from questionable design and manufacturers ignoring the airflow issues of a mounted motherboard. No doubt there was a problem with bad caps at the time, but I've seen Rubycons cook off on this motherboard, too.

The motherboard ('case') temp sensor is way over on the opposite corner of the motherboard and does little more than read the ambient air temp entering from the front vents of most cases. If that sensor had been mounted anywhere near the VRM circuitry, you would probably see significantly higher (like 20+C) case temps then you see now. Of course, this would really just be reading your VRM convection oven's temps in the little tornado near the CPU. I would rather have seen that temp though, then getting a somewhat useless reading of relatively fresh air entering the case.

This is probably just all academic now, but it's still an interesting lesson in cooling and case airflow - at least for primates, anyway. I'm afraid I can't speak for species with opposable thumbs.

The only components you need to worry about is the ones with a heatsink on it.
You might wanna try cementing heatsinks on the regulators, but that may change their values, who knows how touchy this stuff is.
As long as you keep your CPU below 50C, your OK.

I would look at your power suppy, your getting dirty power, thats what the caps filter out is irregular power crap and noise.
Also capacitors are good to 105c, so its not the heat, its the dirty power, probably from overclocking, spurious interference.

[Gibby]

I don't see OST on these caps. In fact it's hard to read anything on them without pulling the caps or pulling the motherboard out of the case. I'm not quite ready to do that yet. But from what I can see, it looks like they have the letters H-M-M, with the last "M" enclosed in parenthesis, like this:

HM(M)

Does that make sense?

Yes - these are Nichicon's HM series. Dell used them on all those Optiplex mobo's that went bad. Nichicon itself isn't a bad brand - they just got burnt by the bogus electrolyte (whose foumula was originally 'stolen' from Rubycon, by some accounts). I would still use HZ-series Nichicons as a replacement without hesitation.

Anyhow, I understand what you're saying about the temperature issue. I thought I was doing a decent job of keeping the temp down inside my case, but after reading what you wrote, I'm not so sure any more.

I never thought about this either until someone else pointed it out. You don't realize there's this little heat whirlpool in that area because you generally look at your motherboard with the case cover removed.

You mentioned a mod that involves exhausting the CPU and Northbridge up and out of the case. Does this mean reversing the fan and cutting a hole in the case?

I wouldn't chop your case. Your replacement caps will probably outive this motherboard even if you don't change anything at all. Simply flipping the CPU fan may help, providing that your CPU temps don't increase as a result. I switched to a heatpipe CPU cooler that exhausted straight toward the back of the case and was high enough to clear most of the I/O connectors.

In the meantime, what about one of those fans that go in an empty PCI slot to pull in more fresh air? Would that be a good short-term solution?

Feel the vent holes on the back of your case by the CPU. If you can feel a decent amount of air blowing out now, then you might not need to push any more air in. The PCI-slot fans are cheap and easy, but don't help out much with the local recirculation of air around the CPU/VRM/northbridge.

I guess it's not enough to just vacuum out the dust every month or so. By the way, the reason you see dust in that first image because I took that picture right when I opened the case and noticed the leaking caps, which was before running the mini-vac. (And yes, I do remove the CPU fan from the heatsink and vacuum it real well too.)

Every month or so? That's admirable. I usually put it off for a few years until I see flames pouring out the back of my case.

...it's obvious that things are steadily getting worse. This might explain why I was able to overclock last year, but not now. As you said, they're probably slowly going out-of-spec. If you're right about the temperature inside the case, then I guess I'm cooking them to death.

Since your motherboard is using the Nichicons, I would lean more towards 'bad caps' as the primary factor. Heat certainly compounds the problem, but there's still tens of thousands of these motherboards running today - in setups similar to yours - that are not cooking off the caps so dramatically.

[cartel]

If you vacuum your cpu be real careful not to rock it or bump it too hard, especially if you have a huge heatsink, you'll crack the cpu real easy.