440MX - my first successful recap

[Oerg866]

Hi.

I noticed for a long time that my noname Geforce 440MX has problems in 3D display, like Trackmania and that the 4 RIFEKING caps are bulging a little. But, it depends of the angle you're looking at it, how it is bulging: not at all, a little, much. If I took a picture it always looks like it's not bulging at all. But I noticed that someone else had problems with RIFEKINGs on Geforce 4s. I got a dead board here and got some SANYO and RUBYCON caps from it and soldered them on the card.



Original RIFEKINGs were 470uF 6.3V. I asked a friend who is an engineer if it is safe to put on these higher ones (1500uF, 2200uF), he said yes.

Result 1: POST like normal, still have to test 3D, but so far it's successful

[c_hegge]

Re: 440MX - my first successful recap

Nice job!

I would advise against changing the capacitance, but if it works, it works.

[Oerg866]

Re: 440MX - my first successful recap

DAMN. The RUBYCON cap is bulging and leaks!!

[PCBONEZ]

Re: 440MX - my first successful recap

You probably have it in backwards or the original cap was bipolar.
or
It's overheating because it's touching that heatsink.

[Oerg866]

Re: 440MX - my first successful recap

The card never heats up at all even when playing games

I went to the bathroom and when I came back I find the screen garbaged (MEMTEST still running fine though) and the cap bulging.

How do I get off that cap correctly without risking it to blow electrolyte in my eyes??

[PCBONEZ]

Re: 440MX - my first successful recap

You are doing okay for someone new to this stuff.
This is meant to be constructive so don't take it the wrong way. - K?

- c_hegge made a good call about changing uF.
- Elaborating on that.

Their is some 'ancient wisdom' that says using more uF makes PC's more stable but as it turns out the 'ancients' were probably wrong.
When they raised the uF they usually inadvertently lowered the ESR because the new caps were either physically bigger or a better grade or both.
- Raising the uF didn't do squat for them, it was lowering the ESR that helped.
-
Another thing is that if ESR matters in a circuit with lytics then the existing uF is usually already over-kill. More uF than needed [and using a higher voltage caps than needed] are both side effects of trying to get the circuit's ESR lower without using more expensive [better grade] caps.
Assuming you don't change the grade of cap, caps with lower ESR are physically bigger caps. Physically bigger caps are either for a higher voltage, are more uF, or both.

There are some circuits on motherboards and in PSUs where we [geeks with soldering irons] know by way of numerous mad scientist experiments [um, I mean experience] that raising the uF doesn't hurt anything, but, extending that ?theme? into other circuits on other parts can get you into trouble.
If the capacitance [uF] affects a frequency or a circuit's timing [sequencing with other circuits] then changing uF can screw things up pretty bad.
-
Most caps on motherboards don't have functions like that but there are some places on motherboards were changing uF is probably a bad idea. An example I can think of is Dell socket 478 and 775 boards for P4 and C2D. Virtually every one I've seen has a 680uF that does something for the north-bridge of the chipset, yet there are several 820uF caps elsewhere on the same boards. It would simplify manufacturing and cut costs to just use another 820uF where that that 680uF is, but they don't. Obviously that 680uF NEEDS to be 680uF. The uF on that one probably affects the chipset's timings. [Sequencing with other circuits.].

All of that said:
-Usually- you get lucky when you raise uF - A LITTLE. [But not -always-.]
The uF tolerance on Lytics is +/- 20%, and uF tends to go down as caps age, so -usually- going up 20% won't hurt anything and as a plus as the cap gets older it gets closer to the original value anyway.
- If you dunno what the cap does it's best to not change uF. Using a better grade with the same uF but with lower ESR is a much better solution.

And, 470uF to 1500uF or 2200uF is a heck of a lot more than 20%!
20% would be from 470uF to a 560uF.

Now, I have no idea what those caps on that card do so >I< can't say if your mod is okay or not. It's just pushing it way past what I would try.
- Did that engineer friend actually analyze the circuit and do the math before telling you that raising the uF that far was okay? I could be wrong but it really sounds like he[or she] perpetuated 'ancient wisdom' without actually looking into it on that particular card.
.

[PCBONEZ]

Re: 440MX - my first successful recap

The card never heats up at all even when playing games

I went to the bathroom and when I came back I find the screen garbaged (MEMTEST still running fine though) and the cap bulging.

How do I get off that cap correctly without risking it to blow electrolyte in my eyes??

Bloated caps don't usually pop from de-soldering.
If you are worried about it just face to solder side of the board [only] while you work on it.
.

[Oerg866]

Re: 440MX - my first successful recap

Thank you for elaborating Yeah the card run stable until it bulged. I found a perfect replacement cap on a MSI board I had laying around.


Laters,
Oerg866

[kc8adu]

Re: 440MX - my first successful recap

watch out as the markings are sometimes reversed .
that got me a while back as i pulled all the bad ones,got distracted,and later ASSumed the white was -.poof!
replaced them again and ok.

[PCBONEZ]

Re: 440MX - my first successful recap

Using used Lytic caps can also be a problem.

Aluminum Electrolytics 'adjust' to the applied voltage within a few hours.

[With aluminum electrolytics]
If you put a 16v rated cap in a circuit with only 3.3v present then after a few hours it becomes a 3.3v cap.

This is because the chemistry inside the cap adjusts the thickness of the aluminum oxide layer [which is the dielectric] according to the applied voltage.
More volts, thicker layer.

If you then put that cap with 16v written on the side but that has been 're-formed' to 3.3v into say a 12v circuit the oxide layer [dielectric] is too thin. Then what amounts to arcing between the plates can occur right through the dielectric. That can result in a cap that's shorted internally pretty fast.
- Doesn't always happen but it can.
If the circuit has some resistance that limits current that can save the cap.

There is a process to restore caps to their original voltage rating.
It's called 're-forming'.
It basically amounts to slowly raising the voltage in small steps [to prevent the arcing while the oxide layer re-forms and thickens] until you get to the rated voltage where you hold it for some time period.
- Everyone that publishes a procedure to do re-forming has a different idea about what times and voltage steps are needed which is why I'm not being specific.
- Most of them work out to between 1 and 2 hours to complete the entire procedure.
[Which indirectly tells you how fast that oxide layer changes in response to applied voltages.]

So, If you dunno the old and new voltages [actual circuit voltages] then using a used cap is not a great idea.
.

[Oerg866]

Re: 440MX - my first successful recap

DAMN. The card works fine now but the problems with textures in Trackmania are still there. I hate this cruel world ;_;

I have another 440mx here so I'll test if that one works with the game.