5 ms·

The art of debugging – or how it took me 3 years to fix my PC

41 points by Reinmar 8y ago

pjc50 8y ago
> This is not a story about CPU soldering or whatever hardware engineers find sexy Guess I should dig out my old writeup of blowing up the floppy drive I needed to boot my PC by dropping a pencil in it, and the ensuing repair..
c22 8y ago
One time I had a tiny magnet inside one of my USB ports. Vexed me for several minutes.
sebazzz 8y ago
I also have a very odd issue which I think has its root in the underlying firmware. I have an Dell Latitude E6520 notebook with a Intel Core i7 Sandy Bridge processor. Though I bought this laptop in 2013, it is still fast and works quite well even with Visual Studio / ReSharper. Only one thing, it is sluggish as hell. Mouse clicks not responding, everything feels slow. Except when I undock it from the docking station and redock it again. Then you feel the fan spinning up, the CPU doing work, and everything becomes smooth. I checked the clock speed, its the same before and after docking. There is nothing keeping the CPU busy either. Its like something is somehow preventing the system from executing efficiently until the system is re-docked. The solution? I don't know, I might never find out. I will just keep redocking my laptop every time I boot Windows.
- amaccuish 8y ago
  Maybe the fans aren't running when they should and the CPU is struggling with heat?
  - Klathmon 8y ago
    Won't Intel chips down clock in that case?
    - billforsternz 8y ago
      I think that's the point.
      - Klathmon 8y ago
        The parent commenter specifically said they checked clockspeed was good.
        billforsternz 8y ago
        Good point, I missed that.
- goldcd 8y ago
  Your real problem is that you have a known workaround - you'll maybe some time trying to work out the issue 'next weekend or the one after', when you've got a bit of free time :) If you can find a bit of time - CPUID & HWMonitor are handy tools for just capturing what your CPU and motherboard are currently up to (compare when slow and when fast)
- free652 8y ago
  I have a feeling you that you got a Nvidia card, so disable Optimus in bios.
  - sebazzz 8y ago
    Yes, that is disabled.
- PascLeRasc 8y ago
  It sounds like you have a slow HDD in there, an SSD would certainly help with sluggishness. I'd also recommend repasting the cooler with Thermal Grizzly paste, it's sub-$10 for enough to do 3 CPUs and it's the best non-conductive paste available.
  - sebazzz 8y ago
    So after redocking my HDD is suddenly fast? No, I don't think so. It is an SSD BTW, I also have a secondary HDD.
- jotm 8y ago
  In my experience, the GPU, USB and SATA are the causes of sluggishness without CPU usage. Not sure what could be the problem with your dock - does it have a USB 3.0 controller that the laptop does not? eSata? You are 100% sure the CPU is not being heavily used by anything after first time docking?
  - sebazzz 8y ago
    The dock is a standard wide-connector Dell port replicator, with indeed USB, DVI connections and possibly eSATA. USB 3.0 did not exist back then.
- abraae 8y ago
  I think yours is a pragmatic attitude in this age of interconnected complex technologies. Debugging things is satisfying but time sucking. My eclipse on fedora has a hideously annoying flashing effect while editing e.g HTML files. Seems to be caused by the top menu disappearing and then appearing again a second later. The janky effect is infuriating, but at the same time, half an hour or so of googling has led to nothing, and I can sense a rabbit hole of frustration if I try and get to the bottom of it. The most pragmatic approach feels to be too just put up with it until something else happens - perhaps a shift to Ubuntu so that I can screen share again will resolve it.
jabberthemutt 8y ago
To taint the clickbait, it of course did not take them 3 years: > I’ve recently been trying to resurrect my old PC. I haven’t used it at all over the last 3 years and moved twice in the meantime
- Reinmar 8y ago
  You're right. It took me over 5 years from the moment I assembled this PC to the moment I found the mistake ;) Anyway, I don't think that anyone would actually think that I worked on this for 3 years. But when applied to software (to which this story lead me), I often hear "it took them 3 years to fix this". It's pretty much the same here. The bug was there, it caused serious issues (also with other systems), I applied various random workarounds and I spent a lot of time on it and yet, it got fixed by an accident much much later.
  - fromthestart 8y ago
    >I don't think that anyone would actually think that I worked on this for 3 years. Your title literally says it took you 3 years. Tragic how clickbait has become so normalized.
    - petercooper 8y ago
      Become? These sorts of title tricks have been common practice in the magazine industry for decades and in newspapers before that. Now we're just in the latest stage where it's been adopted by individuals :-D
  - yesenadam 8y ago
    I don't think that anyone would actually think that I worked on this for 3 years. I 'actually' thought that, of course - it's what it says, or what I read it as saying. (One of your most recent comments says "It took 3 years for 20 people to build CKEditor", which presumably means you worked on it for 3 years, but now I'm kind of uncertain what to believe.)
  - perl4ever 8y ago
belltaco 8y ago
The I/O shields behind the PCs don't fit well, can cut you and are really bad design. Glad that new (premium?) motherboards have them integrated. https://laurentschoice.com/wp-content/uploads/2018/01/io-1.jpg https://laurentschoice.com/wp-content/uploads/2018/01/io-1.j...
canhascodez 8y ago
I went through something similar building my first PC. I had worked a summer job just to be able to afford one, researched all the parts obsessively, and slowly waited for all of the parts to be shipped to rural Alaska. I put the all of the components in the case, flipped the PSU switch, pressed the power button, and -- nothing, not even a BIOS screen. Five red LEDs were lit on the motherboard, indicating a major hardware fault: CPU, motherboard, RAM, or PSU. Well, there was nothing else for it but to start replacing the parts one by one to see what was broken. It would take a day or two to get the Return Merchandise Authorization, and two weeks shipping time in either direction, so every month or so I would get a new part in to try to get this machine to work. That went on for about five months. I was crazed with frustration, and a growing pile of electronics boxes, tools, and testing devices filled the corner of my room. I had a collection of components which I was sure were working: the system at least appeared to boot to BIOS when lying on the workbench, but when all the parts were hooked up inside the case, we got five red lights again: major hardware fault. Finally, at the limits of my frustration, I turned to my brother for aid: "It works on the bench, but not in the box. I don't know why. You figure it out." He returned not five minutes later with the widest grin you can imagine. I was incredulous, and this was a better practical joke than he could ever have devised. He showed me that having the case's reset button (correctly) connected to the motherboard caused the error condition. I was so thankful that I almost didn't want to strangle him!
- seotut2 8y ago
  I've seen situations such as that. Working outside the case, but not in the case is usually points to a short of the motherboard through the case (of course, assuming it is wired correctly).
  - canhascodez 8y ago
    As I recall, it took me a considerable while to determine that the equipment worked outside the case. I think I tried replacing the RAM, CPU, and PSU before doing the mobo, and unless you have an idea that the case itself might be the issue, well, why would you take the rest of the machine apart? And even if you suspected a board short, you might easily be misled into thinking that it was a part of the case touching the mobo inappropriately, and leave the reset/power leads connected: it's a lot easier to punch the case reset button than to use the tiny one built into the board. And of course, figuring out that the system works without the power switch connected is not super useful in itself: that is something of a necessary component. I'm just still amazed that he figured it out that quickly; I really didn't tell him much about what had happened, and I'd hardly had time to leave the room before he'd solved it. What a jerk :)

41 points by Reinmar 8y ago

> This is not a story about CPU soldering or whatever hardware engineers find sexy Guess I should dig out my old writeup of blowing up the floppy drive I needed to boot my PC by dropping a pencil in it, and the ensuing repair..

c22 8y ago

One time I had a tiny magnet inside one of my USB ports. Vexed me for several minutes.

sebazzz 8y ago

I also have a very odd issue which I think has its root in the underlying firmware. I have an Dell Latitude E6520 notebook with a Intel Core i7 Sandy Bridge processor. Though I bought this laptop in 2013, it is still fast and works quite well even with Visual Studio / ReSharper. Only one thing, it is sluggish as hell. Mouse clicks not responding, everything feels slow. Except when I undock it from the docking station and redock it again. Then you feel the fan spinning up, the CPU doing work, and everything becomes smooth. I checked the clock speed, its the same before and after docking. There is nothing keeping the CPU busy either. Its like something is somehow preventing the system from executing efficiently until the system is re-docked. The solution? I don't know, I might never find out. I will just keep redocking my laptop every time I boot Windows.

amaccuish 8y ago
Maybe the fans aren't running when they should and the CPU is struggling with heat?
- Klathmon 8y ago
  Won't Intel chips down clock in that case?
  - billforsternz 8y ago
    I think that's the point.
    - Klathmon 8y ago
      The parent commenter specifically said they checked clockspeed was good.
      - billforsternz 8y ago
        Good point, I missed that.
goldcd 8y ago
Your real problem is that you have a known workaround - you'll maybe some time trying to work out the issue 'next weekend or the one after', when you've got a bit of free time :) If you can find a bit of time - CPUID & HWMonitor are handy tools for just capturing what your CPU and motherboard are currently up to (compare when slow and when fast)
free652 8y ago
I have a feeling you that you got a Nvidia card, so disable Optimus in bios.
- sebazzz 8y ago
  Yes, that is disabled.
PascLeRasc 8y ago
It sounds like you have a slow HDD in there, an SSD would certainly help with sluggishness. I'd also recommend repasting the cooler with Thermal Grizzly paste, it's sub-$10 for enough to do 3 CPUs and it's the best non-conductive paste available.
- sebazzz 8y ago
  So after redocking my HDD is suddenly fast? No, I don't think so. It is an SSD BTW, I also have a secondary HDD.
jotm 8y ago
In my experience, the GPU, USB and SATA are the causes of sluggishness without CPU usage. Not sure what could be the problem with your dock - does it have a USB 3.0 controller that the laptop does not? eSata? You are 100% sure the CPU is not being heavily used by anything after first time docking?
- sebazzz 8y ago
  The dock is a standard wide-connector Dell port replicator, with indeed USB, DVI connections and possibly eSATA. USB 3.0 did not exist back then.
abraae 8y ago
I think yours is a pragmatic attitude in this age of interconnected complex technologies. Debugging things is satisfying but time sucking. My eclipse on fedora has a hideously annoying flashing effect while editing e.g HTML files. Seems to be caused by the top menu disappearing and then appearing again a second later. The janky effect is infuriating, but at the same time, half an hour or so of googling has led to nothing, and I can sense a rabbit hole of frustration if I try and get to the bottom of it. The most pragmatic approach feels to be too just put up with it until something else happens - perhaps a shift to Ubuntu so that I can screen share again will resolve it.

jabberthemutt 8y ago

To taint the clickbait, it of course did not take them 3 years: > I’ve recently been trying to resurrect my old PC. I haven’t used it at all over the last 3 years and moved twice in the meantime

Reinmar 8y ago
You're right. It took me over 5 years from the moment I assembled this PC to the moment I found the mistake ;) Anyway, I don't think that anyone would actually think that I worked on this for 3 years. But when applied to software (to which this story lead me), I often hear "it took them 3 years to fix this". It's pretty much the same here. The bug was there, it caused serious issues (also with other systems), I applied various random workarounds and I spent a lot of time on it and yet, it got fixed by an accident much much later.
- fromthestart 8y ago
  >I don't think that anyone would actually think that I worked on this for 3 years. Your title literally says it took you 3 years. Tragic how clickbait has become so normalized.
  - petercooper 8y ago
    Become? These sorts of title tricks have been common practice in the magazine industry for decades and in newspapers before that. Now we're just in the latest stage where it's been adopted by individuals :-D
- yesenadam 8y ago
  I don't think that anyone would actually think that I worked on this for 3 years. I 'actually' thought that, of course - it's what it says, or what I read it as saying. (One of your most recent comments says "It took 3 years for 20 people to build CKEditor", which presumably means you worked on it for 3 years, but now I'm kind of uncertain what to believe.)
- perl4ever 8y ago

belltaco 8y ago

The I/O shields behind the PCs don't fit well, can cut you and are really bad design. Glad that new (premium?) motherboards have them integrated. https://laurentschoice.com/wp-content/uploads/2018/01/io-1.jpg https://laurentschoice.com/wp-content/uploads/2018/01/io-1.j...

canhascodez 8y ago

I went through something similar building my first PC. I had worked a summer job just to be able to afford one, researched all the parts obsessively, and slowly waited for all of the parts to be shipped to rural Alaska. I put the all of the components in the case, flipped the PSU switch, pressed the power button, and -- nothing, not even a BIOS screen. Five red LEDs were lit on the motherboard, indicating a major hardware fault: CPU, motherboard, RAM, or PSU. Well, there was nothing else for it but to start replacing the parts one by one to see what was broken. It would take a day or two to get the Return Merchandise Authorization, and two weeks shipping time in either direction, so every month or so I would get a new part in to try to get this machine to work. That went on for about five months. I was crazed with frustration, and a growing pile of electronics boxes, tools, and testing devices filled the corner of my room. I had a collection of components which I was sure were working: the system at least appeared to boot to BIOS when lying on the workbench, but when all the parts were hooked up inside the case, we got five red lights again: major hardware fault. Finally, at the limits of my frustration, I turned to my brother for aid: "It works on the bench, but not in the box. I don't know why. You figure it out." He returned not five minutes later with the widest grin you can imagine. I was incredulous, and this was a better practical joke than he could ever have devised. He showed me that having the case's reset button (correctly) connected to the motherboard caused the error condition. I was so thankful that I almost didn't want to strangle him!

seotut2 8y ago
I've seen situations such as that. Working outside the case, but not in the case is usually points to a short of the motherboard through the case (of course, assuming it is wired correctly).
- canhascodez 8y ago
  As I recall, it took me a considerable while to determine that the equipment worked outside the case. I think I tried replacing the RAM, CPU, and PSU before doing the mobo, and unless you have an idea that the case itself might be the issue, well, why would you take the rest of the machine apart? And even if you suspected a board short, you might easily be misled into thinking that it was a part of the case touching the mobo inappropriately, and leave the reset/power leads connected: it's a lot easier to punch the case reset button than to use the tiny one built into the board. And of course, figuring out that the system works without the power switch connected is not super useful in itself: that is something of a necessary component. I'm just still amazed that he figured it out that quickly; I really didn't tell him much about what had happened, and I'd hardly had time to leave the room before he'd solved it. What a jerk :)