Changing voltages.

[ChaosLegionnaire]

Re: Changing voltages.

He said he just wanted the system clean.

define clean. he never mentioned what his version of "clean" was. therefore, u are free to make your own judgements of what "clean" is. if he's not happy, tell him to fix it himself or get another computer! its probably easier for him and everyone else just getting a new computer than to fix the existing one!

The programs weren't opening, the system would take 10 minutes (sometimes 15) to shut down (he timed it) and it would crash. He says it only crashed once, the shutdown thing didn't bother him because he'd just go to bed and the programs weren't loading because they weren't transferred off the hard drive properly (I don't know he means by that). Chrome wouldn't load. Firefox did after a couple reboots. I tried telling him with that many leaking and one bulging, it was just a matter of time before it wouldn't start at all. Now I'm frustrated!

He was calling just about every night talking about how his programs wouldn't load but he's convinced they just weren't properly transferred from the hard drive. I asked what he meant by that and he said something about the old hard drive. He used to have those programs on the old hard drive and because the old hard drive was hooked up to his computer, something about parts of them were on the old drive, parts were on the new drive and he just thinks it wasn't be loaded properly or something. He said if I could just copy the rest of the firefox and chrome over, it'd be fine. I tried telling him we downloaded them from the internet and installed them but I don't think he understands.

looks like your uncle has no bloody clue on what a stable and properly working computer is. i wonder how he put up with that busted system for so long.

but anyway in hindsight, maybe u shouldnt have recapped the whole board since that is quite time consuming. just the busted ones around the vrm would do.

if it comes back to you, tell him the board is too old and has been heavily used, so many of the components are failing. then u can tell him to make a choice. either he just gets a new system and u can help transfer his data over to the new system if he wants u to do it, or u can repair his current system but it will take a week or two or three due to the age of the system and level of wear of the components.

[Spork Schivago]

Re: Changing voltages.

define clean. he never mentioned what his version of "clean" was. therefore, u are free to make your own judgements of what "clean" is. if he's not happy, tell him to fix it himself or get another computer! its probably easier for him and everyone else just getting a new computer than to fix the existing one!

That's how I found the cap. I took the motherboard off, I went down stairs and blew it off with the A/C. I removed the heatsink / fan on the CPU to re-thermal paste it and noticed one was bulging.

looks like your uncle has no bloody clue on what a stable and properly working computer is. i wonder how he put up with that busted system for so long.

but anyway in hindsight, maybe u shouldnt have recapped the whole board since that is quite time consuming. just the busted ones around the vrm would do.

We call him Uncle Al but he's just a friend, not any actual relation. It's just the name he goes by. At first, I just did the five by the CPU on the VRM. It didn't fix the problem so I pulled the other tall ones and replaced them in the VRM circuit. Still didn't fix the problem. I started pulling the other ones and found some that were leaky. I replaced maybe 4 or 5 but decided to try it. I went into the hardware monitor and the voltages and fan speed were stable. When I tested some of the ones I pulled, I noticed some were showing as leaky.

Then I went to do the rest of the board. Without including the VRM circuit, there were a total of 9 that showed as leaky. With so many caps showing as leaky, what do you think actually happened? And how was the computer still running? I almost wonder if me replacing just those 4 or 5 that weren't in the VRM circuit somehow made the other ones become leaky.

if it comes back to you, tell him the board is too old and has been heavily used, so many of the components are failing. then u can tell him to make a choice. either he just gets a new system and u can help transfer his data over to the new system if he wants u to do it, or u can repair his current system but it will take a week or two or three due to the age of the system and level of wear of the components.

Will do! I was thinking if he wants me to fix it again, maybe just telling him to take it to one of those official places. He used to go to Southern Tier PC before he knew I fixed them. He said he took a PC there and just wanted them to remove some of the software and replace the mouse. It took over a freaking month!!! But I think at places like that, it's first come, first serve, as it should be.

[Spork Schivago]

Re: Changing voltages.

Always keep the customer updated on what you find and the progress of their repairs. PCs have become important in day to day life and can't be gone to long without a valid reason.

I always clone a customers hard drive before I do any work. That way if anything goes wrong or something is missed getting copied over it can always be retrieved.

Did you not copy over the Firefox and Chrome profiles from the old PC ? If they are copied over the customer wouldn't even know it had been reinstalled. Even the last session is kept.

Have you run checks on the memory and hard drive ?

Thanks for the reply. I did keep him updated but I think maybe there's where part of the problem came from. For some reason, I seem to struggling converting technical stuff to non-technical stuff. I think when I told him there was one cap that was bulging and needed to be replaced, he heard that part but not the rest about the other caps in the VRM circuit. I haven't tested the RAM or the hard drive yet. I'm waiting on one more cap to come. One of the caps that tests leaky is a 470uF 16v 8mm cap. I've never seen a cap so small in value show as leaky but this little 8mm does for some reason. Is that normal? RAM test will be next and then finally hard drive.

I'm almost tempted to just find a non-leaky electrolytic 470uF 16v around the house and put that in there.

As for the profiles, there wasn't really anything to transfer except for links on the desktop. I don't know why, but he doesn't seem to use bookmarks. I showed him a few times how they worked but there weren't any besides the IE ones that got imported when Firefox got installed. Chrome wasn't on there at all. I had installed that for him by downloading it off the net.

I gave him a loaner laptop to use because I thought it might take more than a day to fix. He's using the loaner.

[dmill89]

Re: Changing voltages.

With so many caps showing as leaky, what do you think actually happened?

Older low-end Asrock board likely loaded with mediocre to sub-par caps (OST is common on entry-level Asrocks) + piece of crap powersupply that was outputting a ton of ripple (well beyond the ATX spec that the motherboard was designed to handle) + time (G31 board, this is now a 7-9 year old system) = lots of failed caps.

And how was the computer still running?

You'd be surprised on how many bad caps a board can have and still "work" (as in at least boot up), but they will be very unstable.

[Spork Schivago]

Re: Changing voltages.

Older low-end Asrock board likely loaded with mediocre to sub-par caps (OST is common on entry-level Asrocks) + piece of crap powersupply that was outputting a ton of ripple (well beyond the ATX spec that the motherboard was designed to handle) + time (G31 board, this is now a 7-9 year old system) = lots of failed caps.




You'd be surprised on how many bad caps a board can have and still "work" (as in at least boot up), but they will be very unstable.

Thank you Dmill89. So the leaky caps needed to be replaced to improve stability but might have allowed the PC to still work, essentially? Does a leaky cap always represent future failure? I know they're not good, but do they always have to be replaced? Is it a sign that they're in the process of failing? Or could they last a couple more years?

[dmill89]

Re: Changing voltages.

Does a leaky cap always represent future failure? I know they're not good, but do they always have to be replaced? Is it a sign that they're in the process of failing? Or could they last a couple more years?

The thing is you never really know, it could fail in a week or remain in its current (degraded, but not completely failed) state for years. If you (or your "customer") plan on keeping the system long term (more than a couple months), it is best practice to replace all suspect capacitors or you will likely end up repairing it again (or replacing it).

[Spork Schivago]

Re: Changing voltages.

The thing is you never really know, it could fail in a week or remain in its current (degraded, but not completely failed) state for years. If you (or your "customer") plan on keeping the system long term (more than a couple months), it is best practice to replace all suspect capacitors or you will likely end up repairing it again (or replacing it).

Thanks. One more question. I noticed over the years my ESR meter has trouble measuring in circuit. I believe I understand why now. I think it could be for various reasons. For example, depending on if the caps are hooked up in series vs in parallel would affect what my meters sees. Since technology is always changing, is there any sured way to measure capacitance and ESR in circuit, regardless of the circuitry, or is it always best just to pull them to measure them? I currently pull them and then measure. I just wanted to know if there's any meters that can somehow successfully and accurately measure ESR / capacitance in circuit.

[Spork Schivago]

Re: Changing voltages.

Nevermind, I think I got my answer. I might be able to successfully get an idea about ESR, but measuring capacitance in-circuit just isn't a good idea.

[ChaosLegionnaire]

Re: Changing voltages.

well, next time when u find faulty components when cleaning a pc, just tell the owner u found faulty components that should be replaced but replacing them could take a week or two. ask the owner for permission if its okay to go ahead to "book" his pc for a couple of weeks to repair otherwise dont bother starting any repairs. just put everything back but tell him if the computer feels "broken" when he gets it back, its not your fault and u did not break anything when u cleaned his pc! its the already busted parts in the pc when u got it!

if its me, i might even make the owner sign a waiver of liability that its not my fault. this is prolly cause of my medical background. e.g. when a patient goes to hospital for treatment and finds out that he has a condition that requires treatment but cant afford the cost of the treatment and asks to be discharged instead, the hospital makes the patient sign a waiver that they are not at fault should the patient's condition worsen or die later on. this ensures u or the hospital cant be sued in court later on! this is called coverin' yo ass!

[Spork Schivago]

Re: Changing voltages.

well, next time when u find faulty components when cleaning a pc, just tell the owner u found faulty components that should be replaced but replacing them could take a week or two. ask the owner for permission if its okay to go ahead to "book" his pc for a couple of weeks to repair otherwise dont bother starting any repairs. just put everything back but tell him if the computer feels "broken" when he gets it back, its not your fault and u did not break anything when u cleaned his pc! its the already busted parts in the pc when u got it!

if its me, i might even make the owner sign a waiver of liability that its not my fault. this is prolly cause of my medical background. e.g. when a patient goes to hospital for treatment and finds out that he has a condition that requires treatment but cant afford the cost of the treatment and asks to be discharged instead, the hospital makes the patient sign a waiver that they are not at fault should the patient's condition worsen or die later on. this ensures u or the hospital cant be sued in court later on! this is called coverin' yo ass!

I like that idea. And the way you worded it, it doesn't sound very complicated at all, you know? Not overly technical or anything. I tried keeping it not too technical but went into capacitors and a little bit about ESR and tolerances. I should have just said I found some faulty components, like you said.

I really thought about doing that waiver thing, with all customers. Just have them sign a piece of paper that covers their ass and my ass as well. If I mess up, I have no problem paying for my mistake. But if I didn't mess up, I don't think I should have to pay. Thanks for the advice!

[stj]

Re: Changing voltages.

the caps probably got taken out by the old psu.

[ChaosLegionnaire]

Re: Changing voltages.

yea. what stj said. did u check the psu as well? i see lots of the pro techs here all check the psu if there are busted caps on the mobo. sry, the thread has gotten too long so i cant remember or find the post that said if u did anything to the psu. if the voltages continue to fluctuate after a full recap, it may be the psu doing that and going bad.

also, something further to add to the waiver, u might wanna attach and take pictures of the faulty component(s) to go along with the waiver so there is physical evidence to support your claim and point of view in court and also to the owner. the owner can have an idea of whats the problem with their computer. so as an example:

I should have just said I found some faulty components

yes and then take pictures of what u found and send it to the owner. then start with the asking permission to "book" his pc for a couple of weeks or so to repair or not.

[stj]

Re: Changing voltages.

no, legally you must keep any parts you remove, and return them to the customer.
no throwing them on the floor / in the bin.

[stj]

Re: Changing voltages.

yea. what stj said. did u check the psu as well?

read the thread!

psu was total shit, and is now replaced by a new SeaSonic. (no fucking about!)
https://www.badcaps.net/forum/showpo...3&postcount=29

[Spork Schivago]

Re: Changing voltages.

yea. what stj said. did u check the psu as well? i see lots of the pro techs here all check the psu if there are busted caps on the mobo. sry, the thread has gotten too long so i cant remember or find the post that said if u did anything to the psu. if the voltages continue to fluctuate after a full recap, it may be the psu doing that and going bad.

also, something further to add to the waiver, u might wanna attach and take pictures of the faulty component(s) to go along with the waiver so there is physical evidence to support your claim and point of view in court and also to the owner. the owner can have an idea of whats the problem with their computer. so as an example:

yes and then take pictures of what u found and send it to the owner. then start with the asking permission to "book" his pc for a couple of weeks or so to repair or not.

The thread did get really long. First thing I ended up doing was checking the PSU and we found the voltages were fluctuating on some of the lines. I learned how to test it with my oscilloscope and saw what I think is referred to as ripple current (it was like a title wave moving across the screen). Once I replaced the PSU, I still had issues. That's when I tore it down to clean it and noticed the bulging polymer looking cap (that turned out to be some electrolytic that looked like a polymer). I'm kind of glad I checked all the other ones now. There were a lot that physically looked good but either had high ESR or showed leaky or the capacitance was out of the 20% range. For the 20% range, it wasn't much over. There where a couple 1,000uF that read around 1,260 or so.

I'm assuming the ESR was high on some because a lot of them read around 0.0something but there were some that actually read like 0.2something and one that even read 0.4something.

[ChaosLegionnaire]

Re: Changing voltages.

wow... thats quite a large range of different failure modes for the caps on that board. i heard that usually electrolytic caps fail by drying up and thus going high esr when their electrolyte dries up and poly caps fail by having higher and higher leakage current until they function like a resistor instead.

since your board had a mix of electrolytic and poly caps, did ALL of the electrolytic caps fail by high esr and ALL of the poly caps fail by being leaky and/or out of range? i'm asking cuz this is useful knowledge/info when trying to analyse the failure mode of different types of caps.

[Spork Schivago]

Re: Changing voltages.

wow... thats quite a large range of different failure modes for the caps on that board. i heard that usually electrolytic caps fail by drying up and thus going high esr when their electrolyte dries up and poly caps fail by having higher and higher leakage current until they function like a resistor instead.

since your board had a mix of electrolytic and poly caps, did ALL of the electrolytic caps fail by high esr and ALL of the poly caps fail by being leaky and/or out of range? i'm asking cuz this is useful knowledge/info when trying to analyse the failure mode of different types of caps.

The two polys actually tested fine with my meter. I had contacted Digikey and got a nice e-mail from them explaining stuff. I wanted to know about these polymer replacements. Essentially, I've read two things. Some sites say you can use them as direct replacements. Other sites say because the ESR is so low, they fire faster and you must use half the capacitance. That didn't really make sense to me though. I went to where I ordered the caps and asked them. This is the e-mail I got from them on the matter:

Hello,
 
I touched based with one of our engineers on your questions, here is what he had to say:
 
Typically low ESR and high ripple current would be what you want from a capacitor so for general use, that would be preferable to a high ESR, low ripple current. 
 
If you're using it in an application with a particular IC, say a voltage regulator output, you don't want to stray from what is recommended to use. ESR is usually more important in this case because regulators have a feedback loop. The location of poles (which cause instability on the output) and zeros (which force the output to zero) of the system are affected by the capacitance and ESR. Normally they work together to set the output as desired. If the poles and zeros are set in the wrong place, the regulator can oscillate or not work as otherwise expected. This is why you don't want to use high ESR caps in the VRM as Ken mentioned. However, some ICs may also need a minimum ESR to maintain stability as well. So this is wholly dependent on the circuitry that the capacitors are going into. 
 
Ripple current is more important for power dissipation. Power dissipated in a cap is I_ripple^2 * ESR. So if he's replacing old electrolytics with polymers, the overall internal power dissipation will be lower. You don't want to go the other way around or you can overheat or blow up the cheap electrolytics. 
 
So in short, replacing caps for ones with lower ESR and higher ripple is advisable in most cases because of the lower power dissipation. However, it can sometimes be an issue if a power IC requires a certain minimum ESR.  

Please contact myself or any of the technical representatives if we can be of further assistance.

This makes sense and it seems in some areas, a specific ESR might actually be required. The electrolytics with varying degrees of ESR and capacitance were all the same make and model number. Unfortunately, I cannot tell you where they were in the board. I didn't record the values. At first, I was pulling one, testing one, replacing it. In the end, I started doing two to four at a time. Pulling them, replacing them and then testing the two to four.

I know the ones by the RAM weren't right at all. High ESR and capacitance out of range.

The ones down by the hard drive weren't right either. There were a bunch of them and maybe by the northbridge (I think it's the NB. The thing with the heatsink).

I should of just recorded the values and the locations. I can't believe I didn't do that. It would have been a great learning experience.

[Spork Schivago]

Re: Changing voltages.

Also, I found an electrolytic to put in this machine (470uF 16v) just to get it up and running until the new polymer comes. I've had it running MemTest86+, doing all the tests. It seems to be stable, running 4GB (2x2GB) DDR2-1066MHz sticks (running at 800MHz) G.Skill F2-8500CL5-2GBPK. Because it's 1066MHz but running at 800MHz, I tightened the timing a bit, 4-4-4-12 instead of 5-5-5-15. I didn't change any voltages or anything because I felt that isn't something I should be doing unless a customer asks me too.

Only way to get the RAM to run at a faster speed is to adjust some other numbers not for RAM. For RAM, my only two options are 667MHz and 800MHz.

It's done 6 full passes and all 6 have passed. I'm going to say the RAM is good. Gonna try booting into the OS now and see if the programs load or crash or what.

[Spork Schivago]

Re: Changing voltages.

Memtest86+ has been running for 5 hours, 24 minutes 10 seconds now. I just checked the temps, CPU Temp - 44 degrees C / 111 degrees F
M/B Temp - 39 degrees C / 102 degrees F.

I noticed there's two case fans. One on the side (glass window) and one on the top (glass window). They both appear to be blowing air out. This doesn't seem right. The top one is by the PSU. The PSU is blowing air out. Shouldn't the top one be blowing air in? It blows the air in, and then the colder air blows threw the PSU which blows it out....am I wrong about this?

What about the side one? I guess that could make sense about blowing the hot air from the CPU fan out...what do you guys think? I don't see any fans in the front. Trying to figure out where's the intake fan. If all three fans (one PSU and two case fans) are blowing out, how is the air getting in? Maybe this had something to do with all the bad caps...

There's two places for 120mm fans, one in the front, one in the back, but no fans there at all. Should I reverse the flow of one of these smaller case fans or try to find a 120mm fan somewheres to add? Maybe even two of them?

[Spork Schivago]

Re: Changing voltages.

I've been doing some research and I think I'm just going to flip the side case fan so it blows in. The top one and the PSU ones should still blow out. Thinking of trying to find a 120mm for the front though and get that blowing in.