If you plan on using NewtLight with a Newton 120 or 130, things are easy. These models can source a minimum current of 35 mA at 5 Volt. In theory it is not possible to fry anything by
drawing a higher current since the chip that is in charge of providing this current (a LTC902, a custom AppleTalk chip made for Apple that was also released commercially as LTC1320)
has a thermal overcurrent protection. According to its data sheet it can even cope with outputs shorted out indefinitely. But I’d rather you didn’t rely on this information, hence you are strongly discouraged from
connecting any hardware that draws more than the 35 mA mentioned above.
The interconnect port of eMates and 2x00 Newtons is quite a different cup of tea. According to the N2 Newton Interconnect Designers Guide (803kB, pdf format), the 5 Volt
power supply is shared among the internal serial slot (a connector on the Newton's main logic board that is e. g. used by the SER-001),
internal flash RAM (during programming) and the LTC line driver. Besides these internal components, the supply can source 500 mA
of additional current, which is shared among the two PCMCIA card slots and the Newton's interconnect port. Any power that is used by the card slots must be subtracted to determine the amount of current available to the interconnect port. However, the maximum current that can be sourced by the interconnect port can never exceed
100 mA.
So far, so good. What this tells us is that for as long as you do not draw more than 100 mA
through a NewtLight adapter that is connected to the interconnect port of a Newton 2x00 or eMate, there’s no risk of frying your machine. The USB light that I used for my tests drew 70 mA
and was thus well within the specification. But...
Former hardware developers are generally curious by nature. Especially so if they do not understand a specification completely...
I had an inkling that as to the specified maximum current of 100 mA , Apple might not have meant "can
never exceed" (as in “the Newton will make sure it won’t, no matter what you try”), but "should
never exceed" (as in “it’s your responsibility that it won’t, and if it does, be prepared for the worst”). In order to clear this ambiguity up, I plugged the NewtLight into a near-dead mainboard (why risk a
perfectly healthy Newton?) and connected a 20 Ohm resistor. This resulted in a 250 mA
current, which proved that my assumption was correct: There is no circuitry inside the Newton that will keep the current through the interconnect port within the specified range.
I do consider it likely that even larger currents won’t create any damage for as long as the total power budget of 500 mA isn’t exceeded, but, unfortunately, this is but an educated guess. Please be
aware that components that are operated outside of their specification do not always die at once. Instead, they might decide to take revenge at a later time when it is much less convenient. So for the time being let
me politely suggest that you refrain from connecting anything that will draw more than 100 mA (0.5 Watt).
If you are the adventurous type (and if you happen to have a spare Newton), try drawing more current for a longer time, and don’t forget to let me know your findings...
By far the safest way to solve this problem would be providing the NewtLight with fuse. However, this would be quite difficult to manufacture and hence advance the sales price significantly, which is
why for the time being production units come without such a fuse.
|