colwell@mfci.UUCP (Robert Colwell) (03/02/90)
I've had an NEC 965 hifi VHS VCR for a few years, and every since it was a baby, it's had a Problem -- every once in a while, during normal operation, it would turn itself off. And while off, it would occasionally wake up, cycle through its lights and displays, run the motor and tape transport a few seconds, and go back to sleep. It would sometimes go completely dark (even the clock display would go out) and it would take repeated AC line disconnects/connects to restore sanity. I bought the schematics for it from NEC and waited for it to enter the Twilight Zone again. Of course, when it did, I'd run to the workshop with it, and it would proceed to work perfectly, with random poking around revealing nothing out of the ordinary. Based on the above manifestations, I was convinced that the micro that controls all the lights and reads the switches was getting random resets, so I connected a lot of wires to various interesting points in the reset circuit and brought them to a terminal strip outside the VCR cabinet. Two nights ago, it went back into Screwup Mode, and I found the problem. NEC did some strange things. The reset circuit is ridiculously complex, with a 5-way diode OR "gate" feeding the "+" input of an opamp. The "-" input had a pair of resistors biasing it to 4.0 volts from the +5V supply. But the "clear" circuit was driven off something called "EVER 5V", and when the machine was screwing up, that turned out to have 4V on it. So the opamp wasn't getting enough voltage on the "+" input to switch the other way and drive out a valid reset. I don't know why NEC thought it reasonable to devote an op-amp and at least a dozen other components to the task of putting out a 5V reset signal. The EVER 5V was drooping because the NPN transistor feeding it was shot. Again, I don't know why NEC didn't just three-terminal regulators to convert the 10V to 5V (there are two transistor-based converters in the 965). Instead they used an old-technology zener/capacitor/bias-resistor with pass transistors. Replacing the offending transistor fixed the machine. I found it very weird that in a high-tech piece of equipment with special IC's everywhere, what should fail but a transistor in the most mundane part of the entire machine. Bob Colwell ..!uunet!mfci!colwell Multiflow Computer or colwell@multiflow.com 31 Business Park Dr. Branford, CT 06405 203-488-6090