Subj : Re: USB lock up - poo! To : Barry Martin From : Ky Moffet Date : Wed Oct 28 2020 13:45:00 BARRY MARTIN wrote: > Hi Ky! > > It is sort of a good news/bad news item. Good news is I don't have to > spend lot of time trying to figure out where the problem is and > possibly what part to replace. Bad news is it is sort of an expensive > item, or pair of items as I'll probably replace with an Intel > motherboard so I can go with an Intel CPU. Unfortunately, it missed the haystack! > KM> Ya know, that's an interesting thought... so you're saying you > KM> can use a larger capacitor in the same spot? Hmmm.... > > Voltage being the larger value I'm thinking about, yes. Usually a > larger voltage capacity also means a larger physical structure which may > not fit in the allocated space. Ah. If I ever get around to attempting such a solder job, I'll have to consider larger caps (plenty of boards and PSUs to cannibalize) cuz now I'm thinkin' that some of the iffy caps of the plague era might not have gone POOF if they weren't being driven to max capacity. > There is also the capacitance (the 'Farad' part, smaller being > microFarad, or mfd, æfd). In DC (direct current) filtering a larger > value (up to a point) will reduce ripple (AC/line noise). again with a > larger physical size. I might go up a notch or two on the capacitance > -- there are other factors like charge and discharge rates, I don't > have the background being 'just' a hobbyist. I'd feel more comfortable > playing with the values in a power supply than in a computer curcuit. > Agree: have seen that done professionally as well as done it myself. That > little trick will get one in trouble with higher speed circuits -- > thinking specifically RF where even the component sitting a little above > the board versus sitting flush will alter overall values. ...I'm not When it's not working at all, I doubt it matters if you fix it so it runs a bit slower than design specs! > quite sure what the capacitors around the Southbridge circuit actually > do: filtering? timing? blocking? I'm thinking DC filter because of > leakage issues so the extra lead length probably wouldn't make any > difference for that short a distance. And can probably halve the feeder > wires plus one as is a common ground. Who knows. But here are some Handy Charts showing what connects to what: https://en.wikipedia.org/wiki/Southbridge_%28computing%29 > KM> Yeah, does seem kinda iffy for that. But as repurposed mine > KM> rarely writes a file other than whatever internal housekeeping > KM> linux does, and if it gets bogified, no one cares. No longer a > KM> critical system, just a handy one. > > OK, yes. I was initially thinking only the USB circuit being munged but > then remembered it is part of the Southbridge circuit, which controls a > whole heck of a lot of other stuff including the hard drives. Back to > not being a good choice for a critical system. Maybe testing/isolation, > though what little testing I do I tend to use the Virtual Linux Machine. Given the above chart, I wonder if a drive attached to PCI-E is out of that loop. > KM> Didn't notice that one, but haven't been paying attention as not > KM> really in the market at present. In fact I haven't finished > KM> moving from Cash to Silver II. Grrr... I hate moving house.... > > Sometimes it's bad enough when isn't moving! Where do you live? I dunno, my house moved. þ RNET 2.10U: ILink: Techware BBS þ Hollywood, Ca þ www.techware2k.com --- QScan/PCB v1.20a / 01-0462 * Origin: ILink: CFBBS | cfbbs.no-ip.com | 856-933-7096 (454:1/1) .