http://pirikukka.blogspot.fiI started my blog when I started to repair Philips Electronic 202 turntable (22 GA 202/33T), then how I bought my own car and its problems that were hidden.
It doesn't get updated too often because I have my part time job, studies and of course fixing my car. And sometimes I think the repair was so trivial I don't even bother make a blog post for it. "Hey I changed the elcap and it worked". Would be fun to make repair videos for stuff that I fix at work but NDA prevents me from talking about our customers' boards, their problems, fixes, tweaks, hacks and of course I am not able to share pictures or videos so that is a bit shame.
Also I need to sometime figure out how I can use blogspot better because default layout is kinda 'meh'.
I would say I am very good 'bug finder' or 'fault finder'. I have always been very curious how devices work internally thus I with intense playing around with devices and software I have found a lot bugs and reported many of them. Sadly too many times I see bug reports, faults and so on to sink at the bottom because many users report same issue over and over again. The problem is nobody ever thinks after market support enough, I have been glad to see many indie developers have found out Mantis Bug Tracker to let community to do the work to organize faults, report them, vote for urgent repairs. Totally awesome system if you ask from me.
Well if you want electronic related answers from me all I know basically is how to fix devices with oscilloscope, multimeter, power supplies, heaters, ovens, repair techniques, professional tools etc.
Basically you can ask me how in industry finding a fault from PCB can be done as quickly as possible, what not to do when you turn your proto board to be manufactured at massive amounts etc.
All faults that need to be fixed might end up costly for you and even if you don't consider something a fault it might end up costing you more than you wanted.
One example is that customer asked us to hand solder battery in place with 6mm distance from one component with 0.1mm margin error. Impossible? No, it is not but surely slows down production and you end up paying it to manufacturer. Then if you suddenly set up weekly quota to double or even triple then if factory ain't flexible then increasing output can be very challenging because one thing is to make extra shift to assemble and test devices but training them takes some time and if board has very challenging condition then many of the units will not meet customer specification and will be reworked to fix the issue. Of course this rework doesn't cost to customer anything but it might hurt your end customer feelings about you if you cannot deliver stuff in time. (Blaming on manufacturer if the quota is changed without a few weeks beforehand is not their fault.)
Another example is PCB recycling where you have added new components to old board, these new components might need to be bent in precision or there are just so many of them. At some point when you're hand soldering 10 - 30 components to your old PCB design it will cost you more than designing new PCB and ordering them after old ones have been used up. In new version there should be slots for new components so wave soldering can be done or even better if you can replace most or all hand soldering components to SMD.
Third example is that your board works perfectly in perfect world but then manufacturer asks you for assistance that they have like 200 faulty boards out of 1000 but test system that you provided doesn't give them enough info how to find which component or components are faulty. So to design a good test system ask yourself: If this component fails then what kind error message it gives? Lets say if you design it badly then the error message is just: Board failed! or Timeout occured! One clever way is to use onboard MCU to 'talk' with other controllers on board and report they are ok. But I have seen an example where error message is just "Timeout". Basically it needs more information than same error message over and over again. Example: "Timeout occurred at U10".If you don't know how to design one yourself then there are companies that have specialized building custom test fixtures. These companies often do superb job designing these custom test fixtures for your board. Remember hand work amount is something that you can control when you design a board! With minimal handwork needed then you're paying mostly just for components and manufacturers own set % profit they want, design poorly and you have extra cost of paying employees wages also with manufacturer % profit included to that.
Then there are repair guys like me whose job is to make sure customer gets their board on time, manufacturer (my employer) doesn't lose too much of his profits or loss is minimal. Sometimes we have to give up on board where fixing it doesn't generate profits or doesn't minimize losses anymore either. My successful repair percentage at factory is above 95% If I am familiar with board already then I can fix it in couple minutes, sometimes needs a hour or two but next time takes much less time because I have become more familiar with board. OF course as customer you do not end up paying for these if your contract says you do not buy faulty units. But then you have to consider if you want to change manufacturer the repair guide they have made or knowledge how to repair them won't come at all or won't be cheap to buy. Well of course I can't say is that a standard contract with manufacturers but where I work our standard contract is that we don't send or sell faulty units to them or charge money for faulty units if the unit will be tested at manufacturer and the faulty units are not customers fault (like change in revision, software or you sent wrong type component info / reel).
As said earlier in I can't say anything specific about our customer boards but in general I can talk about fixing these stuff and in future I hope to update my blog more often.
Home
Help
Search
Login
Register
