e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

[momaka]

Naturally a continuation of this thread… I bought a GeForce 6200 with bad Sacon FZ caps that didn’t have any vents on them! Fun, easy, recap project, even though the video card itself is not worth much. So let’s get down to the pictures/details…

The video card itself looked like this when I got it (click links to see pictures):
https://www.badcaps.net/forum/attach...1&d=1400270134
…yup, those Sacon FZ caps are leaning quite hard due to pushing out their bungs. Here’s a better view:
https://www.badcaps.net/forum/attach...1&d=1400270134
The best part is, I almost didn’t need to desolder these caps in order to remove them. They pushed out their bungs so hard that two of them already had their legs half-way pulled out from the vias on the board. I think I understand what Sacon was trying to do now with these vent-less caps. Basically, as the cap bulges and pushes out its bung, it also breaks away its solder joints and falls off, thus leaving only the “good” caps behind. Brilliant!

We also have us some silver Sacon SZ. These are the bulging caps on the upper-right corner of the video card.
https://www.badcaps.net/forum/attach...1&d=1400270134

I gave the video card a Sacon, be-gone treatment, and here’s the result:
https://www.badcaps.net/forum/attach...1&d=1400270134
Much better… but ugh, they left quite a mess, didn’t they? I tried scrubbing with alcohol (IPA, 99.9%), but that electrolyte must have gotten baked on there or something. It was tough taking it off. Acetone did the trick, though. On a related note, the stock thermal compound on the GPU heatsink was that pink chewing gum crap that wouldn’t wipe with IPA either. Acetone proved to be a good friend here again.

Anyways, back to caps. Two mug shots of the bad boys:
https://www.badcaps.net/forum/attach...1&d=1400270386
https://www.badcaps.net/forum/attach...1&d=1400270386
Looks like a c_hegge style type of execution won’t bring the joyful banging sounds here as they have already released their pressure. But I still have a few “creative” ideas .

And a backside shot:
https://www.badcaps.net/forum/attach...1&d=1400270386

Finally… the recap job:
https://www.badcaps.net/forum/attach...1&d=1400270386

As you can see, I like diversity quite a bit. In that mix of caps, there are some Nichicons, Fujitsu, Sanyo, and Rubycon. To put it in better words, I don’t like to discriminate against caps, as long as they are Japanese and of high quality and reliability . That said, they are all used caps. But I am sure this won’t be a problem as I’ve done this for many of my personal repairs before. I also decided to leave the Lelon OCR caps alone. These are 16 V, 330 uF, 10 mm diameter, and they filter the 12V rail from the PCI-E connector (C17 and C14 on the board).

Some detailed breakdown:

C30, C17, and C14 ---> 12V rail filter. C30 is before a PI coil L3, and C17 and C14 after it. C30 was a Sacon FZ 16 V 470 uF, but as you can see from the picture above, I just removed it and didn’t replace it with anything. I probably will in the future. I just didn’t have any decent 16 V caps in 8 mm diameter on hand. That cap is not all that critical.

Next, C38, C47, and C48 ---> GPU V_core filter (buck type). About 1.33V when powered.
I used 2x Fujitsu FP-CAP RE 4 V, 820 uF and a single Nichicon HZ 6.3 V, 2200 uF. All 10 mm diameter. The Nichicon has a 2005 datecode (H0533), so it may or may not be good. Most are, but a few I had did bulge after sitting un-used. So that’s a bit experimental here, but the Fujitsu’s should take care even if the Nichi went down. 2.5V caps might work fine here too, if anyone ever decides to save such an old card.

C29 ---> GPU Vtt voltage. It’s linearly stepped-down from GPU V_core and is about 1.2V. Just about any cap will do here. I used an SMD Nichicon (what series??) 4 V, 1000 uF. This was a bit of an experiment here as well. I have quite a few of these (pulled from bad PS3 boards) and since I often need 1000 uF in 8 mm diameter for motherboards, I wanted to see if these would work in a pinch for low-voltage rails (such as RAM Vdd and Vtt). I don’t think these are low ESR, but I’m not sure. In any case, GPU Vtt is a linear rail, so it makes no difference what you use here. Many of the nVidia GeFoce 6 and 7 series use this setup.

C23 ---> possibly 12V to 5V step-down DC-DC rail. It’s a buck type. Used a Sanyo WF 16 V 1500 uF from an Xbox 360 12V filter. Never used these before, so this was an experiment too .

C215 ---> 3.3V rail from PCI-E connector. It’s the input filter to the APL1581 linear regulator that feeds the RAM. Originally a Sacon SZ 6.3 V, 1000 uF. Now a Rubycon MFZ 6.3 V 820 uF.

C135 ---> RAM Vdd voltage. Output filter cap for APL1581 regulator. Was a Sacon SZ like above, but I replaced it with another SMD Nichicon 4 V, 1000 uF cap.

After all of this recapping, I stress-tested the video card with some games. It does pretty well with Colin McRae Rally 2004 @ 1024x768 maxed out (minus AA). Portal, however, brought this video card to its knees – that is, the video card worked just fine, but I was getting terrible framerates on certain levels even at 800x600.

Also worth mentioning is that the fan on this video card is quite loud. There is no temperature controller – it runs on full blast at 12 V all the time. Not only that, but the GPU itself actually still runs very hot – 58C idling at desktop and 70C playing games. And that’s with a 24-25C (about 75-76F) room temperature! My guess is either the GPU heatsink is crap (it does have a very dull and rough finish) or my cheap thermal compound is crap. I’m thinking the former, since I’ve used that same thermal compound on other video cards, and they didn’t seem to get that hot… Namely, I’ve done two GeFoce 7600 with it, and both run about 45C idle and 55C stressed. I guess, we will see when I try a copper-base heatsink on it.

Anyways. Overall, I still think this was a fun little recapping project. Good enough for a test PCI-E video card since it has both DVI and VGA outputs. I might donate it to a classmate of mine

[Agent24]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

Nice!

[goodpsusearch]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

Very good job. I like doing experiments with caps too

[c_hegge]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

It's a shame that the sacons were already popped. It would have been fun to blow them all up.

Nice job on the repair.

[Wester547]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

Sanyo WFs are notorious for going bad (a search for them on this forum will affirm that). Not to say you were dumb for using them, you're doing the right thing by using Japanese brands. A very good repair otherwise. Sacons are really, really really bad, but amusing at least for the way they fail...

[goop]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

What are the rubber cap bases for?

Likewise, what are the plastic stands for on SMT caps?

[Agent24]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

They aren't bases, it's just the bung been pushed out from the pressure.

I'm guessing the plastic stands on SMD ones are for keeping the leads in shape, also for showing polarity.

[momaka]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

Thank you all for the kind words.

Likewise, what are the plastic stands for on SMT caps?

To isolate the leads from touching on the metal can of the capacitor and shorting out.
SMD caps are actually just regular radial leaded capacitors. The way they make them is simply by taking a radial leaded cap, putting on a plastic stand on its bottom, and then bending/flattening the leads to the plastic stand.

Because the leads on SMD caps are flattened, you can't unbend them without weakening them too much (usually if you bend them twice, they will break off). So if you want to use an SMD capacitor in place of a regular leaded radial capacitor, you have to solder leads to the pads on the SMD capacitor.

I always save the excess leads from leaded redial capacitors for this and other similar purposes. It's actually very easy to "convert" an SMD component to through-hole. And the reverse, of course, is even easier.

Sanyo WFs are notorious for going bad (a search for them on this forum will affirm that).

Interesting.
I did search and read the threads. If nothing else, they seem to be heat-sensitive at the very least.

That said, I did take like 5 minutes trying to pick which Sanyo WF to use on the video card. Originally, I was going to do it all Fujitsu and Sanyo WF (so the whole video card would have had yellow and green caps), but I noticed that the bungs on some of the Sanyo WFs looked just a tiny bit rounder than on other caps I have. I wasn't sure if that was normal for the Sanyo WFs. Then I recalled that I did see a few bulged in my friend's repair shop on various very old Xbox 360s. Since I got mine from junk 360 boards, there is a possibility that these caps might have gotten heated up quite a bit at some point (a reflow/reball was likely performed on the board prior to me getting it).
So that overall dissuaded me from using them.

I know the Rubycons, Nichicons, and Panasonics I get from the 360s are all troopers. I recapped a crappy ECS P4IBMS motherboard with all Rubycon MFZ and Nichicon HZ caps that were slightly bulged due to being overheated. They all passed ESR and capacitance tests at the time, so I used them. That was more than a year ago, and they are still good and the board still boots.

It's a shame that the sacons were already popped. It would have been fun to blow them all up.

Yeah, it is.
I was thinking about opening them up (which should require almost no effort now ), wetting the paper/foil inside with water, and closing them back up again, then applying power .

I need/want to get my camera working, though. We have an old but decent camcoder that uses MiniDV. I'd really like to put it to good use. The only problem is, I bought a cheap 4-pin to 6-pin Firewire cable that ended up damaging the port in the camera, so now I don't even know if I can get the videos out of it or not. It be great to blow up some Sacons on film! I'm sure it should be similar to an 80's action movie .

[Wester547]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

I did search and read the threads. If nothing else, they seem to be heat-sensitive at the very least.

I'm not knocking Sanyo for those failures. I think once you get low enough ESR wise that lytics, given the augmentation of water content, are bound to be very heat sensitive and possibly unstable (like KZGs and MCZs are)... and probably can't last forever like the older capacitors do (sort of.... I guess they all fail eventually).

As for bulging capacitors in 360s, I'm surprised they don't bulge more often given how hot those bastards run. I'm guessing the bulging ones were in 360s that were ran 24/7 (that's my only explanation besides the towel or oven because 24/7 use precludes the heat cycles that would sooner kill the lead free solder and BGA solder balls... that's the only way I could imagine launch consoles surviving besides seldom use or possibly modding the 360's fans to run at +12V). Still makes no sense to me how often those Rubycon MCZs and Chemi-con KZJs bulge on motherboards that probably aren't as hot as the 360 but how they survive in the 360s.... unless the higher voltage (less heat dissipation) of the VRM input has something to do with it, but that only applies for the VRM input and not the VRM output which has MFZs in some older consoles and those have been observed to have failed on motherboards with enough heat and time....

That said, the GeForce 6200/7600s are both good cards but I think they're haplessly among those affected by the underfill defect.

[goop]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

They aren't bases, it's just the bung been pushed out from the pressure.

Would you say this is an indication of failure?

I have the impression most caps with some space to the board show this.

BTW, SMT/SMD... I don't understand the need for two terms. What's the point? Haven't seen yet THD as counterpart for THT.

[momaka]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

Would you say this is an indication of failure?

10000000% YES!

[QUOTE=goop;448006]I have the impression most caps with some space to the board show this.[/quote
Yes, but it tends to be fairly obvious when they push out through the bottom even a little bit. The bung will look a lot rounder.

BTW, SMT/SMD... I don't understand the need for two terms.

SMD: surface mount device
SMT: surface mount technology
http://en.wikipedia.org/wiki/Surface-mount_technology
I'm sure there is a right way to use those terms, but my fingers are just lazy, so sometimes they press a "T", and sometimes a "D" (whichever is closer). Pretty much they mean the same thing, though.

As for bulging capacitors in 360s, I'm surprised they don't bulge more often given how hot those bastards run.

They do run hot - the CPU and the GPU, that is. Unfortunately, the fans are so innefficient in these (due to restrictive vents) that you just end up with a very hot heatsinks and the components around them barely get blasted with hot air from the heatsinks). Also, the 16V caps are located pretty far behind or to the side of the heatsinks, so they don't get that warm at all.

That said, the GeForce 6200/7600s are both good cards but I think they're haplessly among those affected by the underfill defect.

With good heatsinks, though, they seem to do fine. Only the 7900 IIRC has problems - and to me, it appears that this is caused either because the GPU chip is too big and too close to the PCI-E connector so the BGA cracks, or a combination of the higher power dissipation of the GPU with the underfill defect.

Speaking of cooling, I stress-tested the GeForce 6200 in this thread some more. This time, I used RiveTuner to record the temperature of the GPU. It ramped up steadily from 58C to 69C as I was testing it with games. This means the thermal compound appears to be working fine (as I thought), and just the heatsink is crappy. In comparison, I also stress-tested two Radeon HD2400 XTs. Both still have the pink chewing gum thermal compound. The temperatures were incredibly unstable on one of them - jumping up and down as much as 5C depending on load. Sure the HD2400 video cards have GPU and memory clock steppings (whereas the 6200 is always running its memory and GPU clock on full tilt), but would this be able to cause that much of a change in temperature so quickly? I might change the pink gunk on one of those HD2400 video cards just to see if temperatures improve/stabilize.

[Agent24]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

would you say this is an indication of failure?

10000000% yes!

i have the impression most caps with some space to the board show this.

yes, but it tends to be fairly obvious when they push out through the bottom even a little bit. The bung will look a lot rounder.

Some have a standoff as part of the bung and you will see this normally:


While some have a bung that is recessed:
The cheaper brands usually use the flat, recessed, type, and when these start bulging it becomes obvious.

btw, smt/smd... I don't understand the need for two terms. What's the point? Haven't seen yet thd as counterpart for tht.

SMD = Surface Mount Device and SMT = Surface Mount Technology. SMT refers to the whole manufacturing process and such, while SMD refers only to the components themselves.

[momaka]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

I pity the poor SOB who bougth this one:
http://www.ebay.com/itm/PCI-Express-.../291260627106?

Looks like the same card, doesn't it?
Apparently, the EVGA GeForce 6600LE is almost nearly the same card as the GeForce 6200 in this thread. In fact, from what I gather online, the only difference is the RAM (the 6600LE has 256 MB and it's faster RAM, whereas the 6200 has slower RAM).

Many people stated as being able to unlock another extra 4 pipelines on this 6200 with Riva Tuner. I think I even read one post on a forum where a guy who had the 256 MB version was able to flash 6600 BIOS onto it.

Yeah, I know it's still a crappy video card hardly worth that effort, but I just wanted to post this here in case anyone has one of these sitting and gathering dust. I think it works with other 6200s as well, as long as they have an NV43 GPU, revision A2 (and possibly a few other revisions... I don't remember off top of my head, but it's researcheable). I will definitely be trying to unlock those pipes on mine as well when I get back home for Thanksgiving holidays.

Also attached a picture of the 6600LE from that eBay auction:
https://www.badcaps.net/forum/attach...1&d=1414456989
model number for that video card is 256-P2-N363-TX
...just in case anyone on the net searches for cap issues with their e-VGA 6600LE. All of the info on post #1 above should apply equally to the 6600LE video cards - same voltages on the caps and all.

[momaka]

e-VGA GeForce 6200 [128-P2-N361] - FIXED, SOFTMODDED, and with a new heatsink!

I will definitely be trying to unlock those pipes on mine as well

SUCCESS!

The unlocking of the extra 4 pixel pipes was super simple and easy with Riva Tuner. So the card went from 4 pixel and 3 vertex units to 8 pixel and 3 vertex units. Below are the GPU-Z screens.
Stock with 4p/3v


"Softmodded" to 8p/3v


Note, however, that this is only possible with the NV43 GPU chip and "A" core revisions (and I *think* "B" as well). From what I've read, this doesn't work with the Turbo-Cache version of these video cards (which are also newer, use a different GPU chip name, and built on a smaller nanometer technology).

So I guess the biggest question everyone has is, How much did this "softmod" help performance?
Answer: it depends (well, isn't that always the answer? ).
I tested only two older games: Collin McRae Rally 2004 (CMR 04) and Portal.

In CMR 04, the framerate increased by approximately at least 10 FPS on most stages/tracks (from 40 FPS average and 60-ish FPS tops, to 60 average and 70-80 FPS tops). So overall, the performance gain was considerable. Graphics settings for this game were all maxxed out in a 1024x768 32-bit resolution with AA and V-sync OFF.

However, in Portal (level 18, which appears to be one of the more graphics-demanding levels), the framerate was terrible with and without the softmod, dipping as low as 13 FPS near parts that have that green gooey toxic water. Not sure why this video card struggled with it that much. Could also be a drivers optimization issue. I should probably try HL2 and some other Source games to confirm. Anyways, the graphics settings used for Portal were 1024x768 resolution fullscreen, all settings on high, AF = 2x or 4x, HDR and color correction ON, and AA and V-sync OFF.

Test system specs:
AMD Athlon 64 3200+, socket 939
AsRock 939Dual-SATA2
640 MB of mixed RAM (2 x 256MB and 1 x 128 MB, IIRC?)
Windows XP Pro 32-bit

Next important item I should mention (again) is the GPU temperature. Even though the temperature in our house was now about 70F (close to 19C), that 6200 GPU still ran insanely hot: 55-57C idling on desktop and 65C under full load - which is WAY too hot for my taste. Not only that, but the small fan was loud too. I mean, who wants a mediocre (even for its time) video card that runs loud and hot?

Therefore, I decided to get creative again. Still having tons of Xbox 360 heatsinks at my disposal, I made this:




I guess this steps into the ghetto territory, but let me give a little details on the parts and mechanism. The heatsinks is obviously one of those revised Xbox 360 GPU coolers. The retention bracket/clamp on the back of the card is another Xbox 360 part: the X-clamp (but modified with two of the sides cut). The screws holding the bracket to the heatsink are PS3 case screws. And finally, the modified X-clamp is sitting on a piece of wood (and some vynil shims) that rests on two hot-glue supports on each side. It is made so that all of the force from the X-clamp is spread directly under the GPU. As a result, there is ZERO board flex. That should keep the GPU BGA happy.

And now the best part of this heatsink mod: the temperautre fell down to 48C idle and 55C max under full load - that's nearly a 10C decrease! Moreover, I did NOT need a fan to blow air directly on the heatsink anymore, so no more noise. The only fan near the card was a 120 mm slow-moving fan that I had pointed towards the motherboard to cool off the SB, since it runs hot. So whatever air was moving from that fan is all that was cooling the GPU.

So...
If there is any interest in that pixel pipeline unlocking "softmod", I can post a quick pictorial of how to do it here. I assure you, though - it is very easy! Just enable "bit 0" in Riva Tunner, restart your PC, and BOOM.

[goodpsusearch]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

I really enjoy these projects!

It would be interesting to try something like this with a more powerful GPU. I have PCI-E Nvidia 9x00 (9600 if I remember correct) laying around with a Zalman cooler (http://www.frostytech.com/articlevie...articleid=1982), but the fan has one of its blades broken.

It came to me broken without the fan, so I can't glue it back as I have successfully done in the past and none of my spare fans seems to fit there. So, I need to get creative with it..

[Dan81]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

Sacon FZ,should've known.

I think those are specific to most nVidia GPUs of that era.

Not sure but IMO I think it started with the FX5500. At least my FX5500 was dead when I bought it,with FZ caps puking their guts. I had to replace all of em' with 1997 Rubycon YXGs which came from a old ASUS P2L97. It works now but quite intermittently. It does at times but other times it doesn't. Any advice what I should be looking for? All caps look fine,and the card doesn't get very hot.

[SteveNielsen]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

I really enjoy these projects!

It would be interesting to try something like this with a more powerful GPU. I have PCI-E Nvidia 9x00 (9600 if I remember correct) laying around with a Zalman cooler (http://www.frostytech.com/articlevie...articleid=1982), but the fan has one of its blades broken.

It came to me broken without the fan, so I can't glue it back as I have successfully done in the past and none of my spare fans seems to fit there. So, I need to get creative with it..

What glue and technique did you use for gluing on a broken fan blade when you were successful?

[goodpsusearch]

Re: e-VGA GeForce 6200 [128-P2-N361] – with bad VENT-LESS Sacon FZ… FIXED!

http://www.thepaperplaceco.com/eshop...2456000831.jpg

Made in Japan