Sometimes, a sledge - hammer is the right tool ... :-)

he

d

my

pin

re

s

full

it

I y

ght

the

ic,

oks

of

r

ed

rs

If the power is distributed with actual planes in the PCB it would be difficult to do it any other way. If power is distributed in a 'tree' you could have restricted the current to get a few hundred mV into the board and then look for the Voltage differentials with a meter. OR if its a 'tree', you can start measuring the ohms at each capacitor. With my trusty old Fluke 8060 I correctly identified which hole was shorted to the ground plane of a new PCB. The 2 holes were separated by 0.1 inch ( 2.54 mm. I know YOU know) connected with a 0.01 inch trace on 1 oz copper. Good meters are essential.

G=B2

If it was belching that much heat, you could have seen some of it with a digital camera that's sensitive to infrared. Granted, the typical camera is only sensitive to near IR, and not the far IR that's need for thermal viewing. Lacking a suitable camera, I've used a gun style IR thermometer, with a paper tube over the lens to limit the exposure. It takes a while to scan the board, but it works. I also have several sheets of liquid crystals, that clearly show hot spots if I can get them close to the board.

That's the left coast please.

What I'm seeing is really marginal design. That's the real benefit of computers in electronic equipment design. With modeling and simulation, it's possible to select components that are ever closer to the bitter edge of self destruction. The days of safety factors and over current/voltage/power/whatever overhead are long gone. Everything is designed to blow up should any parameter be exceeded. If the transistor goes into thermal runaway at 100C case temp, then the heat sink will be minimally sized to run it at 99C. It's enough to make me cry, but there's no way to stop the trend.

As for the recession, it's still happening. Home mortgage rates are about 3.75% which is as low as it's ever been and only half a point above the official 3.25% prime rate. My office complex has 3 vacancies out of 6 offices. I'm having to negotiate payment plans with customers. I'm also loaning money to friends, which is really a bad idea. I'm bidding on design contracts against engineers who are willing to work almost for free. My profits are basically supporting my office expenses and taxes. Not good, and no improvements in sight. Doom and gloom.

On the other foot, I'm finding more time to tinker, play, and loaf, which was not the case when I was busy. I'm even getting some exercise and bike rides during working hours. Maybe recessions aren't so bad after all.

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann     AE6KS    831-336-2558

the

my

pin

anywhere

full

it

right

the

schematic,

looks

of

need

years

The likely MLCC and SM failure mechanism

even more likely with the extra 30 deg C or more of PbF soldering

memo to myself, must get one of those IR and sighting laser-spot thermometer guns

As far as I could see, it was a tree. Unfortunately, this thing is quite 'compact and bijou' which, coupled with the fact that the opamps are mostly quite tall SILs, nestling in amongst the rotary and slider pots, makes it difficult to even get the meter probes in anywhere on the business side. Interesting your comment on restricting the current. I did in fact initially put a 150 ohm resistor in the line, but the short was, as it turned out only a few copper cms from the connector carrying the power in, and the short was so complete that all that was left at the first place on the board that you could conveniently measure, was actually only a few mV, let alone a few

*hundred* mV. This made me suspect that the problem was close-in, and I immediately checked all the nearby electrolytics. However, as it turned out, it was a ceramic cap that had failed. That fact alone contributed to to making the job 'less than normal' as I did not for one moment think that it was going to be down to one of them. Yes, of course I've had these caps go both leaky and short, but this one really had gone *so* short that it may as well have been 1mm of wire ... !!

Under normal circumstances, I'm with you on doing the job 'properly', but in today's financial climate, the value of the equipment - not much in this case - and how much the owner has in his pocket, and how much a successful repair at a reasonable cost, is likely to bring him back in the future, are all factors that need to be taken into consideration commercially.

It was a bit of a 'last resort' method, based on those criteria, I freely admit. Still, it was a good feeling to 'win' for a change. I'm sure we've all been there ! :-)

Arfa

Curious - what do you use for a low ohms meter?

I just bought a small non contact IR thermometer, badged Maplin N19FR. It has scan mode, not mentioned in the instructions. Setting on deg F rather than C gives resolution of .1 deg F ,its bigger brothers usually have 1 deg F as best resolution , as higher temp range. Also cone monitoring rather than spot monitoring so probably better for this purpose, but limited to divining to an area of 10mm diameter rather than a particular SMD say. Anyone know of a source of IR transmissive lens ?, glass is totally opaque to room temp objects. Or would a truncated metal cone inverted over the sensing one work well enough as a "macro lens" accessory after homing in to a target area. I am faced with a similar situation. 2x1100W Crest amp that intermittently falesely trips out as brownout when above a few tens of watts of output. Something is loading the + or - 15V rails, like thermal runaway, loading something like exponentially but always recovers , whatever it is. Opamps all over the place and branching rails . I was thinking of defeating the brownout line and let whatever overheat to the point of smoking but will now have a non-destructive extra option when I can clear some 2 foot x 2 foot bench space. Previously I'd tried a PIR pyro sensor but not suitable for this purpose as wavering output for static "image".

anywhere

ohms

full

I use a 1KHz RLC meter Avo B183 , resolution .01 ohm, with 2 diodes across the probe terminals if the 10V pk-pk is excessive, still usable info in a comparative if not absolute sense

LOL I've done the same thing looking for fuse blowing shorts. As far as work goes, I earn earn enough to live with food on the table and a roof over my head. Wish I could get my lazy-ass now 18 yr old son to get any kind of a job, burger flipper, stock-boy, whatever to help out. He's not stupid with an IQ of around 135 just lay-ZEE! Hell when I was his age, even before I couldn't wait to get a job.

--
Live Fast, Die Young and Leave a Pretty Corpse

Found a long-tail pair of TO92 running about 25 deg F higher in one amp than the other amp, while running normally. I will go with a problem in that area to start with. A previous amp had burnt out R and Tr there so presumably a stock fault with leaky transistors, previous were leaky anyway. I doubt if I'd picked up the difference between warm and extra warm by 25 deg just by "digital" testing. Putting a plastic Din plug cover over the pyro with a rubber sleeve holding in place and IR entering the cord hole, narrows the search area but substantially reduces the apparent temp.

I'm not sure "sledgehammer" is quite the right term.

It would be more correct to say that you handed the device sufficient rope to hang itself.

A Bob Parker / Dick Smith ESR/Low ohms meter. Resolves down to a few miliohms. I have used it with some success for this purpose before, but in this case, the results were 'vague' ...

Arfa

My son was the same at that age, but I had enough work to 'employ' him in my business, doing the crap stuff like dismantling and reassembling units, and replacing lasers. I put "employ" in single inverteds, because he never used to surface from his pit until I called him mid morning, and by 2pm, he was burnt out again ...

When work dropped off at the beginning of this recession when all the banks started exploding, I could no longer justify the cost of him. That seemed to do the trick. He went out and got himself a job in a betting shop (yes, they are legal this side of the pond !) and then went on to get a job with the national electricity and gas distribution infrastructure company, where he still is, and is doing very well. Like your boy, he is very intelligent, and a computer genius when he puts his mind to it. It's just that he was, as your boy is, very lazy. Once he's home from work, it's still hard to get him to move from the couch, but at least he goes every day, and is on time, so maybe there's hope for your lad too .... :-)

Arfa

I wonder if there's anything biological in this. I'm not the only person who's observed that, when visiting my parents, I feel as if all the energy has been drained out of me.

It was the right term for me. I was trying to conjour the image of a piece of delicate electronic equipment, dealt with (successfully) by the use of a bludgeoning and completely inappropriate 'tool' ...

Probably one of those east - west humour things again.

Arfa

We're having some fun, so don't anybody get too serious...

I don't see what was inappropriate about the tool, nor do I consider what you did "bludgeoning". If anything, it was quite the opposite -- not unlike handing a man with a leg injury a cane so he could stand up. Once he's standing, you can more-easily locate the injury.

Well, normally, 'appropriate' tools for this type of fault diagnosis, would be considered to be meters and 'scopes and small soldering irons, and so on. A bloody great power supply with a current limit set only to protect itself, seemed to me both 'inappropriate' in the normal range of tools, and 'bludgeoning' in that it was capable of doing serious physical damage to tracks and vias, in the process of trying to persuade something to declare its faultiness by releasing its magic smoke.

All I sought to show, along with an amusing image, was that sometimes, thinking outside of the box that would normally be taught to students in college, and using what less experienced people would probably think of as outrageously inappropriate tools and methods, can result in the 'cure' side of a 'kill or cure' equation.

I do, however, take your meaning, William

Arfa

Might want to try using a higher end DMM which supports four wire resistance measurements. I currently use a 6.5 digit Keithley 2000 which supports this, and has a resolution of 100 micro ohms. It's worked pretty well for me finding low resistance shorts, but they're a bit pricey on the used market. An HP 3478A 5.5 digit DMM can be had for $125 USD or so, and has the same 100 micro ohm resolution as well as 4 wire support.

The problem with using a power supply is that after the short blows there may be a sudden voltage overshoot as the supply recovers. Here at work I just hooked a DSO up to a variable power supply set to 5 volts with a current limit of 3 amps driving a short. I removed the short a number of times and observed that there were spikes greater than 35 volts which lasted several microseconds. I would think that there'd be some potential to blow some more sensitive circuitry. I would use a power supply as a last resort because of this.

Yes, agreed. It was a method of last resort in this case, as the value of the item, and what it's owner would be prepared to pay, precluded putting an awful lot more labour time into a problem that many years experience of working on similar equipment, told me was going to be tricky and time-consuming to locate, using 'traditional' methods ...

In terms of overshoot from the short blowing itself off, I set the current limit at 1 amp in the hope that this would allow enough current to get the faulty item smoking, without actually blowing it back open again.

Arfa