PDA

View Full Version : piggybacking chips. Universally useful?



tezza
February 2nd, 2008, 11:45 AM
Flushed with confidence after de-soldering and replacing chips in three of my computers I've going to have a crack at replacing the (suspected) faulty RS-432 transmit chip in my BBC (b) micro. I've order the part.

One question though. Piggybacking (simply by friction) a new chip onto a suspected faulty one for diagnostic purposes seems to work for RAM chips and it certainly worked when I was diagnosing the problem with my Osborne 1. Can this method be used for other chips though (e.g. my RS 423 one)? I suspect with more complicated ICs this method won't normally work as it depend on exactly what was wrong with the chip? It might even damage the new chip?

Am I right in assuming this, or is this diagnostic method more universal than I think?

Sharkonwheels
February 2nd, 2008, 05:56 PM
The way i've done it is the past, is use REALLY SMALL diagonal cutters, and
cut the Vcc leg of the bad IC, as HIGH as possible, and bend it under so it won;t get power, and then solder the new one on top, bridging to the remaining lower leg on the PCB.

T

Druid6900
February 2nd, 2008, 06:45 PM
No, it's not universal and it only works if the chip is suffering from an open, not a short.

I haven't, personally tried it with anything other than Xk x 1 DIP RAM, so, I can't say for sure.

Would I do it? Possibly, if I had a whole drawer full of that particular chip and all the chips around it. I'd rather spend the time removing and replacing a suspect chip than risk it with only one know good chip on hand.

modem7
February 2nd, 2008, 06:45 PM
Am I right in assuming this
Correct.

A very simple example. Faulty chip is incorrectly driving output pin 5 low. You piggyback (effectively putting in parallel) a good chip and it correctly drives pin 5 high.
Either one chip will win or pin 5 will end up at neither a high or low state (sitting at a voltage in between high and low).

So how did the RAM trick that you used work?

I'll keep this simple because you can research this on the Internet. There are various buses on the motherboard, and each bus has multiple chips
with their outputs connected to the buss. Only one device on the bus is allowed to output (drive high or low) to the bussat any given time,
otherwise the situation I covered in the first sentence will happen.

And so chips with outputs on a bus have "tri state logic" outputs - either a high state, a low state, or a 'high impedance' state. When a chip is not
outputting to the buss, it's outputs go into the 'high impedance' state. From the buses perspective, it is almost like the chip isn't there.

And so your faulty RAM chip (outputs to the data bus) must have failed in a way that resulted in its outputs always being in the 'high impedance' state.

But your RAM chip may have failed in such a way that the 'piggyback' technique you used would not have worked. For example, the failure may have resulted
in the second data pin always being driven low (like the situation described in the first sentence).

tezza
February 3rd, 2008, 12:39 AM
My question has been answered. Thanks.

nige the hippy
February 3rd, 2008, 03:43 AM
And so your faulty RAM chip (outputs to the data bus) must have failed in a way that resulted in its outputs always being in the 'high impedance' state.

But your RAM chip may have failed in such a way that the 'piggyback' technique you used would not have worked. For example, the failure may have resulted
in the second data pin always being driven low (like the situation described in the first sentence).

In which case it will tend to get hot, as it's output will always be fighting with outputs from other chips. in which case you will have already spotted it!

It's not a guaranteed fix, but it's another technique.

I would tend to shy away from the technique for other chips, they tend not to have tri-state output stages, therefore more of an even fight.

SwedaGuy
February 3rd, 2008, 09:05 AM
Oh, we had to do this kind of thing years ago for a kitchen VDU controller. The company that made it decided that they wanted to use this peripheral with a newer system, and thus would require more RAM.

I don't remember all the details (though I could pull the schematics if anyone is interested), but we had to bend out the CS pin, solder on a short piece of wire and solder the other end to a pad on the board. I think there was another pin we had to reroute as well, but I'd have to look to remember which one.

I was amazed at the time that these contraptions worked at all...

But they were prone to overheating...