On Monday 13 January 2003 18:29, somebody wrote > Declan, you can hardly call that a technique at all, or do you ? > Smile, but still, I had a lot of fun reading, you vocabulary and > expressions always make me laugh, it's quite entertaining. > Greetings
Thank you Friedel. No, measuring the voltage on the gates of Fets can hardly be called a technique. I was losing 10mA, and couldn't trace that. I asked for ideas. I didn't have to resort to some of the more exotic suggestions received, that's all I was saying. I was reccomended everything from a high gain FET amplifier with special probe, to an electron microscope. It appears to be a static issue. -- Regards, Declan Moriarty. > > At 08:44 a.m. 13/01/03 -0800, you wrote: > >Well, I have the board sorted. I left it lie over the pagan > > holidays, and approached it again in the new year. A number of > > things didn't survive the hibernation, and I gave it overvoltage > > which took out a number more. Bulk replacements followed, and I was > > left with these sort of problems again. > > > >This board had 40 class B output stages from a floating 1.0V supply > >with fets, and was running off batteries. It transpired that all > > faulty cells could be eliminated with these two methods. > > > >1. Monitoring the 1.0V supply & the supply current on no load, and > >switching every stage on one at a time. Either the voltage or supply > >current would move on a dud fet pair > > > >2. Checking the voltage on the drivers. Some would go high perfectly > >and not go low; Others might go low perfectly and not high; A few > > did neither. > > > >I ended up swapping a lot of fets, and a couple of driver latches. > > One driver latch took 10 mA extra when one particular stage was > > switched high, although the output stage was perfect. I imagine > > that was the original complaint, but can't prove it. After I > > replaced the fets twice or three times, I finally went for the > > chip. > > > >-- > >Regards, > > > >Declan Moriarty. > > > >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > >On Friday 27 December 2002 18:54, somebody wrote > > > > > The fact that "some work and some don't" would support an > > > oscillation or RFI problem. > > > > > > Your comments reminded me a problem we had years ago with a video > > > crosspoint switch. Small DMOS FETs were used as switches. They > > > were packaged in small metal (TO-18) cans. The boards built in > > > engineering worked fine but the units done by the assembly > > > department were completely dead. The FETs were shipped with > > > small conductive rubber "O" rings around the leads to protect the > > > FETs from static. The assembler thought they were mounting pads > > > and left them attached. When the FET was pushed down against the > > > board, the "O" rings were not apparent. I'd hate to tell you how > > > many engineering hours it took to find this very simple problem. > > > > > > The lesson here is when you hear hoof beats, check for horses > > > before you go looking for zebras. > > > > > > Fred Townsend > > > > > > Declan Moriarty wrote: > > > > On Wednesday 25 December 2002 06:08, somebody wrote > > > > > > > > > Two thoughts come to mind. > > > > > > > > > > 1) Might the circuit somehow be oscillating? I have seen > > > > > circuits of this kind that would sing. The problem went > > > > > unnoticed because applying a meter or scope probe added > > > > > enough capacitance to kill the oscillation so everything > > > > > looked normal. The fix was to apply a nf or so of capacitance > > > > > from the gates to ground. > > > > > > > > These fets are directly driven by latches. There is smoothing > > > > in abundance > > > > > > > > > 2) Could this be a RFI issue? Are there strong fields > > > > > around, such as a broadcast station in the neighborhood, that > > > > > the circuit is picking up and causing the FETs to turn on? > > > > > > > > I hardly think so. 2 of these work, this one works but loses > > > > 10mA. > > > > > > > > > Fred Townsend > > > > > > > > > > Declan Moriarty wrote: > > > > > > Here's one to tease yourselves over. I have a board > > > > > > to fix and for once, I do not know how to go about it. > > > > > > > > > > > > I designed the thing. It's battery powered, and is > > > > > > consuming batteries. There is a 1.0V line with nearly 40 > > > > > > class B fet stages, using logic level fets (Irlml6401 > > > > > > p-channel on top and irlml 2502 n-channel underneath) The > > > > > > gates are tied; > > > > > > > > > > > > That stage is pulling 10 mA on no load. It should draw > > > > > > basically nothing. It draws 10 mA even with the load > > > > > > disconnected. The 'Off' condition has -1.5V on the gates; > > > > > > the p channel fet sees the voltage from +1.0 to -1.5V and > > > > > > is fully on, and the n channel reverse biased. 'On' it is > > > > > > the reverse - the p channel is reverse biased, and the n > > > > > > channel sees +2.5V. > > > > > > > > > > > > Now the tracks are too thick to use a millivolt meter and > > > > > > trace where the 10mA is going. 10 mA doesn't show - I had > > > > > > difficulty tracing 100mA. How do I find the $*�"! current > > > > > > leak?? > > > > > > > > > > > > I had one thing - a fet test program which applies logic 0 > > > > > > to all gates and will drive one at a time high, as I press > > > > > > buttons. I couldn't find any gate leakage. The 5V will > > > > > > usually leak onto the 1.0V if a gate is gone, and I can > > > > > > find the fet pair that way. I will go around the fet pairs > > > > > > with a voltmeter and check the values, but expect to find > > > > > > nothing. -- Author: Declan Moriarty INET: [EMAIL PROTECTED] Fat City Network Services -- 858-538-5051 http://www.fatcity.com San Diego, California -- Mailing list and web hosting services --------------------------------------------------------------------- To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB CHIPDIR-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
