CPU with broken SMD capacitor

[Jirka1111]

Hi to everyone. This is my first post, so please be forgiving.

I have Intel Core i5 7600K CPU with one SMD capacitor broken off. It happened during transport. I still have tha broken SMD capacitor, so possibly it can be soldered back to place, but the problem is, it also broke off with the tiny piece of copper pads. In one side, there is still left a tinny little fragment of it, so maybe it can work. On the other side it is unfortunately completely missing.

But I was wondering, if those capacitors are in parallel, maybe I can solder it to another pad of the same node. Unfortunately I don't know. I am just too afraid to use ohmmetter to figure out which pads are on the same node.

I was trying to find some schematics of Intel Core processors, but didn't find anything.
Can anyone please help me with this? Do you thing using ohmmeter will damage the CPU?

[dicky96]

Re: CPU with broken SMD capacitor

Does the CPU work?

[momaka]

Re: CPU with broken SMD capacitor

I wouldn't worry about it too much.

I've used CPUs and GPUs like this with 1 or 2 busted caps. That's not to say these capacitors are not important - they are. However, most computer hardware can tolerate a certain "overhead". After all, this is why it's possible to overclock CPUs and GPUs: the manufacturer uses slightly lower speed than the absolute maximum the CPU/GPU can sustain, so that way they can guarantee that a whole batch of the same binned CPUs will work without crashing.

So with that broken cap, most likely the noise floor on the CPU V_core will go up slightly. If you're into overclocking you *may* find that the CPU can no longer overclock as high as it did before (though, with only 1 missing cap will likely not produce any noticeable results most of the time.)

Therefore, I second what dicky96 said: try out the CPU and see if it works. I'm willing to bet it will, unless some of other capacitors have sustained damaged too - a cracked ceramic cap can easily become short-circuited, and that will prevent the CPU from working. On that note, it's probably also best NOT to re-use that broken off capacitor, unless you have a meter and can verify that it is OK and not shorted.

Do you thing using ohmmeter will damage the CPU?

Nope, pretty sure it won't.
Multimeters can't supply enough current when testing, so the voltage they pump out when measuring resistance will drop, especially since CPUs tend to have low resistance to ground.

[Jirka1111]

Re: CPU with broken SMD capacitor

OK, Thank you for your replay.
I will use ohmmeter to find another node on the chip to be able to solder the capacitor back and I'll post any results if successful.

[momaka]

Re: CPU with broken SMD capacitor

... to find another node on the chip to be able to solder the capacitor back

Don't!

Unless you can check that cap for proper capacity and that it's not short-circuit, don't use it.

As I stated above, ceramic caps can become short-circuited from too much physical force. So you may end up putting a shorted cap on your CPU, and that would surely not work.

It's safer to try the CPU without that capacitor. If the CPU crashes or is unstable, then worry about adding that cap in. Running the CPU without that cap will *not* destroy it or hurt in any way.

[Jirka1111]

Re: CPU with broken SMD capacitor

OK, so here is the update. I took another undamaged i5 7600K CPU and using Ohmmeter I found out that the two capacitors were actually in parallel configuration.
I measured the one that broke off using ESR Capacitance meter. It measured the capacitor correctly and showed its capacity value to be 10uF.
So I just decided to solder it back on the CPU on top of the other one. It is a little bit off the center but I didn't want to mess with it any more. I was also afraid about soldering iron temperature. I usually solder with iron set to 320..340°C, but in the case of CPU soldering I lowered the temperature to 280°C, so it was quite difficult to get the tin melted properly.
I tested the CPU putting it through some extensive stress workload while monitoring temperature and voltages and it passed flawlessly.
So that's it.
Thanks guys for your thoughts.

[dicky96]

Re: CPU with broken SMD capacitor

That was a very inventive way to fit the broken capacitor .

I will keep this in mind for any time I see a CPU or BGA with a capacitor broken off, and the pad(s) missing. Well done and thanks for the tip!

Rich

[Jirka1111]

Re: CPU with broken SMD capacitor

My only concern is about the electrical current being conducted through the connecting copper path. I mean, the total amount of electrical energy transmitted to and from these two capacitors used to be carried through the two separate parallel copper paths. But now the same amount of energy, each time the two capacitors are charged or discharged is being transmitted though only one copper path, instead of two.
Given the voltage value in this case to be just circa 1.5Volts and quite large capacity of the two capacitors (2x 10uF) I am afraid the only one thin copper connecting path is being stretched to the limits.

But I am not sure about this, it's just my guess. Maybe someone with enough knowledge in this field can do the math..

[ChaosLegionnaire]

Re: CPU with broken SMD capacitor

yeah im also not confident this fix will last because stacking two smd mlccs on top of one another increases the mechanical stress on the package of the mlccs. also, i've read that excessive solder fillet height can also lead to increased mechanical stress on mlccs. over time, this stress will cause the mlcc to crack and crumble, resulting in a dead short. this may fry the cpu tho in most cases, u may get lucky when the mobo vrm and/or the psu detects a short and trips the system power, saving the cpu and maybe mobo cpu vrm from being fried.

[Jirka1111]

Re: CPU with broken SMD capacitor

I'm sorry, but isn't it quite in contradiction? If the capacitor cracks, how can it create circuit shortage? Wouldn't it create circuit break (open circuit) instead?
I have never in my life came across a shorted SMD ceramic capacitor.
I've seen a few capacitors with dry joints as a result of extensive mechanical stress and thermal cycling, but never shorted one..

[momaka]

Re: CPU with broken SMD capacitor

My only concern is about the electrical current being conducted through the connecting copper path. I mean, the total amount of electrical energy transmitted to and from these two capacitors used to be carried through the two separate parallel copper paths. But now the same amount of energy, each time the two capacitors are charged or discharged is being transmitted though only one copper path, instead of two.

That's not going to be an issue here... ever.

Yes, these capacitors have extremely low ESR and can shunt many Amps of current in an instant... but averaged out, the current isn't that high and nowhere the limit of what these traces can carry. The only reason they don't stack the MLCCs and instead place each on their own pad is for manufacturing reasons - stacking MLCCs is hard for machines and more likely to lead to cracked/damaged parts... not to mention the issues that could arise trying to do that with wave soldering.

So again, fear not.

I was also afraid about soldering iron temperature. I usually solder with iron set to 320..340°C, but in the case of CPU soldering I lowered the temperature to 280°C, so it was quite difficult to get the tin melted properly.

Unless you keep your soldering iron on the CPU for like 10+ minutes, the temperature won't really matter that much. These modern CPUs have fairly thick Copper heatspreaders on the other side on their CPU cores... so they've got a good deal of thermal mass on the CPU core. Thus, to overheat the core, you'd really need to be abusing the living ghosts out of that CPU with your soldering iron for a long time for anything to happen to it.

yeah im also not confident this fix will last because stacking two smd mlccs on top of one another increases the mechanical stress on the package of the mlccs.

Not really - not when hand-soldered like in this instance.
Now, if a machine was to do it, like I mentioned above, it could easily become a manufacturing nightmare with parts cracking if the pick-n-place SMD machine isn't perfectly dialed in to install the parts on top of each other... which is why you NEVER see SMDs installed on top of each other from the factory by a machine. But doing it by hand, it's not a problem at all.

also, i've read that excessive solder fillet height can also lead to increased mechanical stress on mlccs.

If you've read this in a technical article, could you please link to it? I haven't been able to find anything on that matter, so I don't think it's something to worry about. I also have a ton of old hardware with excessive solder on the SMD caps, and none have failed. More often that not, MLCCs fail due to board flexing and vibration from external forces on the device (i.e. user dropping or bumping or flexing the device.)

this may fry the cpu tho in most cases, u may get lucky when the mobo vrm and/or the psu detects a short and trips the system power, saving the cpu and maybe mobo cpu vrm from being fried.

Unless it's a really badly-designed or very old motherboard with a single-sided VRM (think Pentium or earlier or those cheapo ECS/PCchips boards), most CPU VRMs will detect the short-circuit and shut down. CPU will certainly not die - after all, a short-circuit will cause the CPU to just not get power.
.
.
Honestly, fellas, you're overthinking this one waaaayyyyy too much.

A single missing MLCC or two will not prevent the CPU from working in like 99.999% of the cases probably. At worst, CPU *may* be unstable.

Just last week I was messing with a salvaged socket 775 CPU. It had no less than 3 MLCCs chopped off. One was just chipped, but still on the CPU. I checked it, and it read shorted. Same MLCC did not read shorted on another CPU with the same specs. So I just took my pliers and ripped the shorted MLCC out (didn't bother trying to solder it, because I've tried that before, and it's a PITA and adds MORE RISK to f**k up the CPU from getting solder on the pins/pads.) After that, the CPU had 4 missing MLCCs.... which I couldn't care less to replace. Instead, I YOLO'd the CPU in a junk motherboard and ran it through some CPU stress-testing (Intel Linpack). As I expected, the CPU passed no problems.

I have never in my life came across a shorted SMD ceramic capacitor.

I've seen a good number of them on LCD monitor T-con boards, for whatever reason. Also, neighbor gave me a 5.1 receiver to fix a few years back after they moved it and it magically stopped working after that. Issue was shorted ceramic SMD caps - probably got cracked when the receiver got moved, as they were close to the middle of the board and there wasn't much support for the PCB there.

[Jirka1111]

Re: CPU with broken SMD capacitor

Thanks Momaka for your comments, I appreciate any shared opinions.