Logic analyzers.

[aman]

Back in '83 at Rockwell I had a 32 lead logic analyzer for diagnosing bad boards. It had a resolution of 5 nanoseconds.

I could see PAL inputs and outputs and logic chips and even had clips for the microproccessors. Often trouble was just a short between two pins or a solder void on one. The tool was invaluable and it was from HP for 50 grand.

Whats available now and what kind of resolution are you getting? Just curious.

Just aman

[cheryl]

The stuff that I know about ranges from the Agilent 54621D Mixed Signal Oscilloscope (have you tried one of these yet? omg...it works so smoothly with the oscilloscope part and SO much easier to use than a big ole logic analyzer) with a bandwidth of 60 MHz, 2 analog channels and 16 digital chanels with 25 ps of graticule resolution for about $4k to the Agilent 1680AD, 136 channels, 800 MHz with a time tag resolution of 4 ns for about half of what you paid for back then.

 

Obviously, we purchase a lot of stuff from Agilent Technologies.

[aman]

Jeez, it's almost getting affordable for home labs. Do you find the problems on the boards are still mostly bad soldering or bad chips.

In '84 I used to have a ratio of about 10 bad solders, 3 bad chips, and 1 bad multilayer board short in diagnosing bad production for QA of computer interface boards. It was fun every day. New puzzles to solve and each was worth 10 thousand to the company.

And by the way, glad to meet you Cheryl.

Just aman

[cheryl]

Greetings, Aman,

 

The problems i tend to see varies by the vendor/manufacturer of the board, and the cost. Typically, the higher the cost of the board, the better the type of test the board is put through before being released.

 

On the older equipment, assuming no user error (blown fuses, fried resistors...or my favorite..."the voltage leaked out of the capacitor as smoke" ) i tend to see problems due to the frequent handling of equipment...BNC or N connectors bent, often where they connect to the board, encoders that have worn out, dim displays, loosened jumpers... In a couple of cases, we have even worn out the memory simply from the number of times the unit has been calibrated (you can only read/write to it so many times after all).

 

Most often...i see failures with the power supplies. They tend to be OEM'd by the equipment manufacturers rather than produced in-house, and i don't know if the quality isn't up to snuff, or if small changes made in the manufacture of the power supplies that "shouldn't" make any difference indeed do.

 

Then...on the newer equipment, i see failures that i suspect have to do with either bad solder, slightly misaligned components, or hairline fractures of the boards, that got worsened during shipping. It seems, though, that doing self-repair or component-level repair is going away quickly...so with these failures, especially while the unit is under warranty, we don't spend a lot of time troubleshooting. Even if a site has the ability to do surface mount soldering (i'm fortunate to have great eyesight and a very steady hand), these new chips...even if they can be purchased, require special hoods or mountings to solder them to the board without damaging them in the process.

 

So...any more...we seem to troubleshoot down to the module, then check a couple obvious things (specifically fuses and connections) and consider ourselves lucky if we can get ahold of a new module.

 

It has forced us to look at the repair strategy for a company. In some cases, we have selected more expensive products because they gave a longer warranty, had a better repair strategy, were easier to work with, or were more reliable. And...with some instruments...the extended warranty option turned out to be worth its weight in gold. All depends upon how much the equipment gets abused.

 

:help: ~blush~ Sorry...i can tend to ramble on a bit...

[jvanhalderen]

I got a SEGA MEGADRIVE!!! 16 Mbit!!!! 32 f/s Woohooo!!!

:lame:

[aman]

I used to debug recently engineered boards for pre-production.

The problem I usually ran into was too few capacitor to ground connections. The engineers always thought that a single ground path for several chips could handle a bunch of loads without creating too much noise to interfere with at least one of the transistors. Some of our first boards were garbage.

I hope the PALs are better quality now. I had about 95% programming success and had to replace quite a few. It might have been the programmer we used. It was a Hewlitt Packard too.

What I liked about the logic analyzer was you could see the sequence of events on the circuit board as it actually happened. I guess that was why I kept getting the raises. You have to see the voltage levels on the board in your head and what they are doing and where they ae going. You kind of have to shrink your brain to hyperspeed and feel the board in your head. Go where no man or woman has gone before. :cool:

I guess all the new stuff is surface mount or socket. Do you still have the old hot air, fry the board surface mount repair?

I know solder suckers are good for surface mount caps and diodes but chips could be awfully messy.

Me Jabberjawin too,

Just aman

[cheryl]

The old zit-poppers? ~laughing~ We still have a few of those around, but don't use them too often. Its mainly solder tweezers for the caps and resistors and an iron for the connectors and encoders (thank goodness those are still through-hole...the longevity and durability just seems to be better). The shop does the more sophisticated soldering stuff...i just do stuff from my bench.

[aman]

I got some stuff we subcontracted to an assembler. 50 small boards with about ten through hole and three surface mount chips besides caps. I pulled 19 out that failed first time. I looked at the solder joints and couldn't believe what I saw.

The contractor was contacted and the cheif solderer came to my lab. I had a board set up under my stereo microscope as an example of the problem. The solderer had never even seen a stereo scope. The soldering had been done by eyeball. No wondr I had so many problems.

The moral is, never accept the lowest bid.

Just aman

[cheryl]

~gasping~ At least get a sample of their work first.

[aman]

Hey Cheryl,

Are you allowed to do any work as union or do you have to call in a union solderer to come in or are your free? At Rockwell I had to call the union to send over a tech at $15/hr to solder any repair work that I found.

They usually sent me a really incompetant tech that had already worked 8 hours so we got charged time + 1/2 for crappy work.

Once we had some batteries for the GPS system ready to be launched into space sitting on the loading dock. No union guy was available so the head of our dept. drove them over since they would be ruined sitting, and our dept. was fined $20 grand.

The batteries were $100 grand apiece. The union strangled us.

Can you imagine that the GPS system in space could have been held up and hundreds of thousands lost because no union driver was available.

I finally had to contract a union tech full time to sit around and if I found a simple problem, I had to take the board over to him and point it out and ask him to fix it. Then I'd inspect his work, give it back and ask him to fix it right. It sucked.

Just aman

[cheryl]

Ewwww....that sounds awful!!

 

Naaahhhh..... When i was working on systems for space research with the university, we were so small we were able to do the soldering ourselves. i was never a part of the actual launch so i don't know what was required once the team actually started to work with NASA.

 

In my job these days...since we are actually a manufacturing/production company, we BETTER know how to solder and solder well. The design, test, and experimental stuff is all done at our benches. The actual production stuff is done on the production floors...but all of those people are employees of the company. i have never heard of any union issues, disputes, dues, or anything. It is possible that the people we hire are members of some union....but i can't imagine i would have worked with these people for 8 years and not have heard any mention of union issues....especially in this environment when we have had to lay people off.

[aman]

Howdy Cheryl,

Have you ever had the chance to hard wire a breadboard? It's tedious and meticulous work but fun since it's R&D. The A6 to G27 and Y41 to G26. Count and attach over and over. I did one board for a week and when I checked power to ground I found a short. I tried to find it counting over the whole wiring system again and couldn't find it. Finally I hooked the bare board to a power supply and sloly crancked up the voltage until I saw a tiny wisp of smoke. It had been a tiny piece of wire debris, one piece of a tiny multi-strand wire that I couldn't see under my mess of wires. The short was gone and when the chips were put in the board worked almost perfectly but now the problems belonged to the engineer who designed it. Fun stuff.

Just aman

[cheryl]

~grinning~ Oh yeah...they still teach that in school. Not the troubleshooting technique You used, though...that's great! In fact, when recently coaching Sciece Olympiad, we used breadboards to mock up circuits for "battery buggy" (little battery powered cars) and also for creating a pattern of wires, chips, leds, and such for one student to write instructions, and a second student to follow and try to create the same thing.

 

i never had to follow the instructions of another, but rather got to mock up the circuits myself...so i had a lot more flexibility. In my breadboards used at Science Olympiad now, i still have my pre-cut, pre-bent jumper wires for connecting grounds and power. i tried to lay those out first so that they laid down flat upon the board and didn't get in the way of the other wires that i would need to fiddle with more often in my design.

 

~grinning~ The guys loved to work with me. Besides the fact that it was easier to troubleshoot, see, and change the design on a "neatly" wired breadboard...the pieces tended to stay in a LOT better. On some of the boards, they had been used and reused so many times that keeping the wires in where they were supposed to be was a big plus!

[aman]

I also had to hardwire a BUS to be able to plug the boards into so it could talk to our IBM PC. The protocals were fun and I don't imagine they changed much but it was fun. Building our own boards and talking to them on screen and with a keyboard is a great experience. I worked with good people except I have no idea how it happened but our group got a manager transfer and the guy had his masters in agriculture.

I ordered a new power supply for my HP analyzer when the original bit the dust, (3 grand) and it arrived rush delivery so I set it on my desk and planned on installing it early the next morning before the engineers arrived. My AGRi manager stayed late, put it in himself and almost burned up the whole 50 grand machine.

I was a little upset the next morning. I hope you work with great people but mainly good bosses.

Just aman

[cheryl]

~laughing softly~ Now the computer stuff... i try to only know enough to be dangerous...

 

i'm fortunate, that most of the time my bosses are really good mentors as well. So no matter what else is going on, they're doing their best to look out for me, giving me good opportunities and making certain i work at the things that are keeping me from getting further more quickly. When they're doing that good of a job looking out for you, they tend not to screw up on the other things, like working with management and getting the resources we need...

 

But....its just a good family there. As dysfunctional as it can be at times with a bunch of geeky engineers.

[aman]

What kind of voltage protocols are you working with? We had three main ones since we developed interfaces between military and PC or larger systems. We had 0 volts to +5, 0 to +10, -5 to 0, and -15 to 0, with the higher voltage being binary 1. All the systems were basically similiar with only different voltages but the -15 to 0 of one of the Mil. systems was less affected by noise and the most dependable.

Do you get to assemble cables? I hated that. Boring. I like R&D.

Just aman.