MSI RS482M4-ILD (MS-7191) motherboard recap

[momaka]

I guess I might as well keep the motherboard recap threads on a roll. I have done quite a few recaps over the years, taken pics with intention to posting them, and then never did. So now is the time, I suppose. Like I said, I hope to revive the motherboard forums a little bit with these recaps.
…

So… shortly after recapping the ECS MCP61SM-GM seen in this thread, I decided that it was finally time I get a motherboard with a PCI-E slot. I did have an AsRock 939Dual-SATA2 that had it already, but it was a finicky and unstable motherboard, so I wanted something better and more simple.

In no time, I found a perfect candidate on eBay: an MSI RS482M4-ILD (MS-7191) motherboard with just a few clearly-visible bad caps that was sold For Parts or Repair. $25 later ($14 for mobo + $11 S&H), and I got it.

Here is what it looked like when it came:






Teapo caps on a motherboard? – Who thought that was a good idea?! Luckly, the CPU VRM output was all done with Panasonic FJ.

As usual, I whipped out a cap map for the motherboard and got to cappin' them Teapos. So here is that:
https://www.badcaps.net/forum/attach...1&d=1498861749

The cap rails and quantity of caps on them can all be found in the above image. As such, I am only going to post the cap changes I made for each rail below.

#1: CPU 12V VRM input
No changes made here. Even though the caps were Nichicon HM with early 2005 date codes, I decided to leave them in and see how far they make it as I've seen many other HMs with early 2005 date codes still working. I don't actually remember if I pulled one of them to test on an ESR meter or not. But if I left them there, they were probably fine.

#2: CPU V_core / VRM output / low side
No changes to the original caps. However, there were 4 free 8 mm diameter spots. Three of those I filled with Rubycon MFZ, 6.3 V, 820 uF caps (from Xbox 360 motherboards).

#3: Northbridge Vcc / VRM output / low side
Another buck-regulated rail, so the original OST RLX caps had to go. Caps I used:
1x Rubycon MFZ, 6.3 V, 820 uF, 8 mm Ø
1x Rubycon MFZ, 6.3 V, 2700 uF, 10 mm Ø
This made the total capacitance slightly lower than with the OST RLX caps. But the overall ESR should more than make up for it (22 mOhm for each OST RLX vs. 7 mOhm for the MFZ).

#4: RAM V_dimm / NB Vcc
I am actually not sure if this rail is for the RAM DIMMs or the Northbrdige (or both), as there is another 2.5V rail on the board. But with three caps on it, I imagine it is both important and provides a moderate amount of current. So it probably is for the RAM DIMMs at the very least. I used:
2x Rubycon MBZ, 6.3 V, 1800 uF, 8 mm Ø
1x Rubycon MFZ, 6.3 V, 820 uF, 8 mm Ø
Perhaps a bit overkill for a linearly-regulated rail (both in terms of total capacitance and ESR), but that may have something to do with the caps I had in stock at the time.

#5: RAM Vtt (termination voltage)
Another linear rail, generated by U9 on the board. For that one, I used a different mix:
1x Rubycon MBZ, 6.3 V, 1800 uF, 8 mm Ø
1x Panasonic FJS, 6.3 V, 820 uF, 8 mm Ø
1x Nichicon HM, 6.3 V, 1000 uF, 8 mm Ø
More overkill…

#6: Northbridge 5V VRM input / high side
Changed only one of the KZG caps on this rail with:
1x Rubycon MBZ, 6.3 V, 1800 uF, 8 mm Ø
In retrospect, I think I should have used another of the Rubycon MBZ caps from #5 above on this rail to replace the other KZG. Perhaps that will be work for some rainy day.

#7: USB 5V filter for the front and rear USB ports
No change. The two original Chemicon KZG 10 V, 1000 uF caps were left as is.

#8: 5VSB rail filter
Only changed the Teapo SM cap with the leftover Chemicon KZG from #6 above. Yes, I know this is not exactly a good replacement, but I have more faith in 16V KZG caps than I do in 8 mm diameter Teapo SC and SM series. And this is not a high-stress rail.

#9: 3.3V STBY
This is another linear rail. It is always ON and generated from the PSU's 5VSB. Replaced its lone Teapo SM 1000 uF cap with a Nichicon HM 1000 uF cap. If I had more entry-level low-ESR caps like Chemicon KY, I would have used them just about everywhere for these linear rails. There is absolutely no need for the ultra-low ESR caps I used above.

#10: Northbridge and/or RAM V_dimm secondary 2.5V rail
No changes. It came with a Panasonic FJ cap from the factory. I guess this must also be a very important rail if MSI thought about using a Panasonic cap here. I've seen this motherboard on eBay a few times now, and it seems that this spot is always populated with a good Japanese cap.

#11: Southbridge Vcc
Another linear rail. Out with the Teapo SM 1000 uF can and in with a Rubycon MBZ 1800 uF cap.

That is all. I did not fill the empty cap spots near the PCI/PCI-E slots. I thought about re-using some of the Teapo SM that weren't bulged, but then decided against that, knowing how Teapo SM is likely to bulge and leak.

Pictures of the board after the recap (sorry for the low quality):
https://www.badcaps.net/forum/attach...1&d=1498861749

And a beautiful red PCB finish:
https://www.badcaps.net/forum/attach...1&d=1498861749

As far as stability: I have had ZERO issues with this motherboard since the recap. It has been in use in my secondary PC (which I also use mostly for playing old games) since May of 2013. So it has seen a little over 4 years of use now.

It came with an Athlon 64 3000+ CPU and a stock heatsink. I upgraded the CPU to a 3500+ sometime in late 2015 or early 2016 or so. And finally this last February or March (2017), I upgraded the CPU again to an FX-57 (the fastest socket 939 single core CPU ). For those of you who remember, it is actually this one from the ghetto mod thread. What a beast!

This is also the motherboard that also opened the eBay “flood gates” for me, so to speak. Ever since I got it, I've been buying crap off of eBay with bad caps and recapping it just for fun (as some of may have noticed with all of the Sacon FZ video cards I have posted here). On that note, I have probably spent more money on these parts than I would if I had built a brand new PC with mid/high-end components. However, I don't think a new PC would have been as much fun as fixing these old motherboards and video cards. Although some of them were simple recaps, I have still learned a great deal about repairs with just about every motherboard and video card I did. So overall, I think it was well worth it.

[Wester547]

Re: MSI RS482M4-ILD (MS-7191) motherboard recap

Teapo El Cheapo strikes again!

Good work overall.

Teapo caps on a motherboard? – Who thought that was a good idea?! Luckly, the CPU VRM output was all done with Panasonic FJ.

Teapo caps in a PSU = bad idea too, as far as I'm concerned.

#1: CPU 12V VRM input
No changes made here. Even though the caps were Nichicon HM with early 2005 date codes, I decided to leave them in and see how far they make it as I’ve seen many other HMs with early 2005 date codes still working. I don’t actually remember if I pulled one of them to test on an ESR meter or not. But if I left them there, they were probably fine.

The only failed 2005-dated HM I have seen so far have been in the VRM output or general +3.3V / +5V filtering capacitors. Haven't seen a +12V VRM input 2005 HM go bad yet, but even the 2001-2004 16V HM used on the +12V VRM input seldom failed compared to the VRM output / +5V / +3.3V filtering capacitor failures. I say keep them.

#6: Northbridge 5V VRM input / high side
Changed only one of the KZG caps on this rail with:
1x Rubycon MBZ, 6.3 V, 1800 uF, 8 mm Ø
In retrospect, I think I should have used another of the Rubycon MBZ caps from #5 above on this rail to replace the other KZG. Perhaps that will be work for some rainy day.

Looks like the original capacitors were 1000uF 16V NCC KZG. I don't understand why they didn't use 6.3V or 10V KZGs on the +5V input.

#7: USB 5V filter for the front and rear USB ports
No change. The two original Chemicon KZG 10 V, 1000 uF caps were left as is.

Not sure if keeping KZG there or Teapo SZ on the PCI-Express +12V input filter is the best idea... but time will tell I suppose.

#8: 5VSB rail filter
Only changed the Teapo SM cap with the leftover Chemicon KZG from #6 above. Yes, I know this is not exactly a good replacement, but I have more faith in 16V KZG caps than I do in 8 mm diameter Teapo SC and SM series. And this is not a high-stress rail.

KZG don't need much of any stress to go. But I have yet to see a NOS 8mmx20mm 1000uF 16V KZG bulge. That doesn't mean it can't happen, of course - I think if the 6.3V KZGs can do it then the 10V and 16V KZGs can too. IMO Teapo SC is a replace-on-sight brand/series regardless of case diameter.

That is all. I did not fill the empty cap spots near the PCI/PCI-E slots. I thought about re-using some of the Teapo SM that weren’t bulged, but then decided against that, knowing how Teapo SM is likely to bulge and leak.

Yep, Teapo SM and SZ are terrible. Right down there with CapXon/Fuhjyyu/GSC.

Pictures of the board after the recap (sorry for the low quality):
https://www.badcaps.net/forum/attach...1&d=1498861749

I've seen WAY worse.

[Uranium-235]

Re: MSI RS482M4-ILD (MS-7191) motherboard recap

now its only a matter of time till the ATI NB deballs

[momaka]

Re: MSI RS482M4-ILD (MS-7191) motherboard recap

now its only a matter of time till the ATI NB deballs

True.

The NB on that motherboard does run extremely hot if you don't point a fan at it and if the case ventilation is not great (like on many OEM PCs).

I have this motherboard in my current secondary/gaming PC. I used to have an 80 mm fan pointed at the motherboard and thought that was good enough. But when I upgraded the GPU, clearly that 80 mm wasn't cutting it, as even the winter temperatures for my CPU were a bit high. That's when I changed the 80 mm fan for a 120 mm medium-speed fan (again, pointed at the motherboard). It keeps the SB cool and some of the air also reaches the NB and keeps it at a reasonable temperature... or so I thought. Turns out, the NB runs very, very hot in the summer now (my room reaches 30°C sometimes), especially when I play games. I always thought that keeping my case slightly open would keep the temperatures in check (and indeed it did - for the CPU, that is). But to have the NB actually run at a reasonable temperature, I have to keep the case side almost 1/2 way open. So all of these years, I have been running the motherboard with the chipset pretty hot. But I guess even that was better than whatever computer this motherboard was in before I got it. If left without any direct airflow (like most OEM builds are), the NB heatsink will run scorching hot. With my setup, even at its worst, I can still keep a finger on the NB HS (would guesstimate the NB core is running around 55°C tops).

So if I am lucky, that MSI motherboard should still have plenty of life left in it.

Also, between this ATI RS482 NB chip and nVidia nForce 6100/6150, I'll take the RS482 any day. The nForce 6100/6150 are even worse when it comes to failures. And then there are the nForce 7 and 8 series of chipsets - pray if you have one of those and want to see it last (that means I should too, as I have recently acquired an EVGA 750i-SLI motherboard in "working" condition).

Teapo El Cheapo strikes again!

Teapo strikes all the time.

More interesting would be to see an old Teapo strike and still be in spec.

Looks like the original capacitors were 1000uF 16V NCC KZG. I don't understand why they didn't use 6.3V or 10V KZGs on the +5V input.

Probably to get lower ESR and higher ripple current specs.
I don't mind what they did at all, because 16V KZGs seem to be a lot less failure-prone than their 6.3V counter parts (save for the 16V, 470 uF KZG - those are bad too).

Not sure if keeping KZG there or Teapo SZ on the PCI-Express +12V input filter is the best idea... but time will tell I suppose.

Neither are a good idea, indeed . But I couldn't care less. Most video cards have their own filter caps, anyways.
Perhaps when I get some small 16V, 470 uF caps, I will replace it. Right now, all I have for 16V, 470 uF are tall Nichicon PJ/PS caps - about twice as tall as that SZ, so that's a no-go, because of getting in the way of GPU.

I suppose when I buy some smaller 6.3 V, 1000 uF caps, I will replace those tall Panasonic FLs and Rubycon MFZ/MBZ/MCZ caps around the memory slots too. It really is a waste putting such good caps on linear rails. I prefer to use them on buck-regulated rails, as they have very good ESR and ripple current specs. I also prefer to use them when they are near a hot regulator, as they are taller and have bigger area to dissipate the heat from the regulator.

[Wester547]

Re: MSI RS482M4-ILD (MS-7191) motherboard recap

Teapo strikes all the time.

More interesting would be to see an old Teapo strike and still be in spec.

They could remain in-spec for a short while after bulging if the solvent and solution on the electrolytic paper is still somewhat wet. But that would soon change once the ruptured can's pressure relief vent opens, leading to leaking electrolyte and complete dry-outs.

Probably to get lower ESR and higher ripple current specs.
I don't mind what they did at all, because 16V KZGs seem to be a lot less failure-prone than their 6.3V counter parts (save for the 16V, 470 uF KZG - those are bad too).

They could have used 1500uF 10V, 1500uF 6.3V, or 1800uF 6.3V KZGs, all in 8x20 case size, and retain the same ripple current and ESR specs. I think they were just using what they had on the assembly line/spooler to save time, I.E laziness.

Yup, the 470uF 16V KZG are terrible. They are the only 16V-rated KZG I've seen bulge in storage. That said, I've seen many other kinds of 16V-rated KZG (and KZJ) bulge or go-out-of-spec under moderate or less heat. So none are trustworthy.

Neither are a good idea, indeed . But I couldn't care less. Most video cards have their own filter caps, anyways.

If those I/O filtering KZG go out-of-spec, the USB and other peripheral devices may not work correctly.

I suppose when I buy some smaller 6.3 V, 1000 uF caps, I will replace those tall Panasonic FLs and Rubycon MFZ/MBZ/MCZ caps around the memory slots too. It really is a waste putting such good caps on linear rails. I prefer to use them on buck-regulated rails, as they have very good ESR and ripple current specs. I also prefer to use them when they are near a hot regulator, as they are taller and have bigger area to dissipate the heat from the regulator.

I would not put the weak MCZ or MFZ series near hot linear regulators. Neither series handles heat well. MBZ would last somewhat longer. FJ and FL would probably outlast all of them, but none of the ultra-low-Z series by any of the Japanese majors fare as well as their higher ESR next-of-kin.

[Maykon Cavaleiro]

I have one of these on the bench, I found it in a landfill, someone had removed the capacitors and changed the values and your post helped me a lot, knowing which capacitors they were. I'll post the photos soon