Philips 90s remote control - resetting

[Buriedcode]

Hi,

I bought a couple of "vintage" (~1995) remote controls with jog wheels - partly because the jog wheel assembly is useful for control surfaces and partly out of interest about that weird period in the 90's when remote controls got more features that were never really useful.

So, although its old and second hand it looks pristine inside, and no silicone oil leaking from the rubber keys - everything looks new, with the exception of some severe corrosion on a couple of pads on the board.  It has a built in clock, that, when set, does keep time accurately all whilst drawing about 30uA standby power. 

However, about 70% of the time, when almost any button is pressed (seems to be some buttons more than others) it resets - that is, the little LCD freezes, and after 2-5 seconds, it resets and powers up with flashing clock - which it does when the batteries are first put in, flashing the clock so it can be set.

Whilst most remotes are pretty simple - micro, carbon fingers for buttons, IR LED(s) and a decoupling cap - this has a fair bit more for driving the LCD and a higher power siwtching circuit for two IR LEDs.
It has two oscillators - a 32kHz RTC, and a 4MHz xtal/resonator, that, according to the micros datasheet, switches between.

I have checked the micros reset line/WDT line, and it does go low for ~10us, but this is *after* the reset of the micro, so it looks like the watchdog timer fires during power up, rather than causing the reset.

I'm stumped really. The only things I can think of are:

1) SMD aluminium capacitors.  Theres only two on the board, a 47uF bulk across the batteries, and a 2.2uF used on the reset line for a POR RC circuit.  These both have corrosion on their pads, the only corrosion on the board which leads me to believe they've leaked, but measuring them in circuit they both measure in spec.

2) Right next to the 2.2uF cap is a small ground trace that connects to the capacitors for the 4MHz xtal osc.  That trace is discoloured and corroded, but still appears to have continuity to ground.  Perhaps its affecting the 4MHz oscillator, which I assume, the micro switches to when a button is pressed.  A brief probing of the XTALOUT pin driving that crystal does show 4MHz (well, 3.998MHz) briefly before it resets.

I realise this is cheap and somewhat useless equipment, but I thought I'd ask here if anyone has any ideas/thoughts about it.  It's mid 90's stuff, with a 4-bit NEC microcontroller so I'm not sure if its prone to ESD issues, or whether just age.  It's not important to me at all, I just can't let things be broken without at least trying to fix them 

[Haenk]

I assume the caps - when this springs into action, it draws a bit more current - if the caps are dry, that might be too much, voltage breaks down, reset.

[Buriedcode]

I assume the caps - when this springs into action, it draws a bit more current - if the caps are dry, that might be too much, voltage breaks down, reset.

Yup, that was the problem.  First hint was the only corroded pads were the ones on the two SMD electrolytic caps - I assume from leaking, and corrosion moving down the caps leads.  The second was the reseting.  Weirdly, overnight, after being powered for 12 hours, it worked OK, no resets - I'm assuming that the decoupling cap was so bad it needed a few hours of sitting at DC to do its job.  Constantly taking the batteries out to disassemble then reassembling didn't give the cap enough time to recover to a point where it was just above the reset threshold.

Before subjecting the board to any kind of soldering stress I scoped the power supply right at the battery terminals to confirm - pretty big (>0.6V) dip in power for ~10us, from 2.95V down to 2.3v, and back up to ~2.8 (whilst the IR LED's are active) .  There was also significant overshoot when the LED's were turned off, again, about 0.6V.

Testing showed the IC reset at 2.2-2.3V so right on the threshold.

So yeah, twisted them off, cleaned pads, replaced with a decent cap, now there is pretty much no undershoot or over shoot, just a ~0.15V dip in power when the IR LEDs are active. Works fine now.

I left this thread just in case someone else has similar issues.  And for equipment thats pushing 30 years old, especially media equipment, be sceptical of those SMD electrolytics.

[tooki]

I left this thread just in case someone else has similar issues.  And for equipment thats pushing 30 years old, especially media equipment, be sceptical of those SMD electrolytics.

Any equipment that old is bound to have iffy caps, whether SMD or through-hole.

Just yesterday at work, someone brought me a failed HP 54510B oscilloscope that won’t boot any more. I found at least 14 leaky caps in the power supply, so I ordered replacements for all of the caps in the PSU. And they were all Chemi-Con; I shudder to think how bad it would be if HP had used cheap caps!