Posted By
TLC
on 2009-06-05
16:34:58
|  Re: Why is the TED chip so vulnerable to frying?
ESD is not that problematic, as speaking of the TED (or any other C= chips). First, these chips are manufactured in NMOS... NMOS is not as prone to ESD as CMOS chips of the time have all been. Second, ESD can be a source of problems only if you expose the chips to it somehow (say, in order to give it an ESD shock, you somehow have to touch the pins of the chip first) -- this is rarely the case here, the only possiblitity being crude ways of connecting/disconnecting peripherals, touching the pins of the user port or things like that, unless you had disassembled the computer first... In reality, as long as you don't do something nasty, you don't really have much to worry about ESD.
Periodic dilatation of the silicon chip (because of heat-up and cool-down) is usually the problem (...not especially dilatation, but hairline cracks on the chip, that start up from defects in the crystalline structure, and develop by each heat-up / cool-down cycles... these hairline cracks would reach vital traces or FETs sooner or later). I remember reading some articles about that... this is a very old one:
http://www.zimmers.net/anonftp/pub/cbm/documents/repair/heatsink.txt
As for using a PC fan or something like that...
You can usually feed such fans from some volts below their nominal supply voltage. But that takes some experimenting. For example, it's been a well known trick to use "large" (8 cm) fans and operate them from 7v instead of 12v amongst PC modders (back then, when today's regulated fans were unavailable), to replace noisy high-rev fans of CPU heatsinks. This trick might be applicable here (not the large fan, but underpowering an otherwise high-rev 12v PC fan). ...You also need much smaller air flow here than PC fans are originally designed for... the heat dissipation of these chips is in the order of below 1 Watt(s). If you take a look at the schematics of the Plus/4 power supply, you can notice that _everything_ (excluding the datasette and the 9v AC supply of the User Port, but including everything else) is powered off from a 7805 voltage regulator... whose current is (by definition) limited to 1 Amps. Multiple 5 Volts by 1 Amps -- you'll find that the _whole computer_, including cartridges, must (by definition) dissipate _below_ 5 Watts. ...Have a compare of that value to todays PC components... 
Also, keep in mind that fans operating above something like 1000-1500 rpms "are" noisy... especially if you have to listen to it directly (which seems to be the case here). At least it won't be absolutely quiet anymore.
If needed, small fans could be operated from the unregulated rectified voltage (+9v unreg), which is about 9-10v. You can draw this from some point at C7, C8, C9 (+9v unreg, as seen at the bottom of this page: http://zimmers.net/anonftp/pub/cbm/schematics/computers/plus4/plus4-310164-2of4.gif ). I'd probably go for that option first anyway, as this is a separate power circuit from the main one, and is only used for the datassette and the 9v AC output of the User Port. (Including a variable series resistor to the pack, with which one could set a reasonable speed, would be the best).
It's interesting to read that most of you met dying TEDs instead of CPUs... my experience has so far been the opposite (...say, usually the CPU would give it up first).
|
