Quote:
Originally posted by unregistered
pardon if I've not been paying attention:
is pHaestus picking up an analogue signal to input to the Maxim ?
or has it already been digitized ?
"Replacing the Maxim IC with a homebrew setup is a possibility, but it would be very expensive by comparison."
ballpark # ?
I think I am confused
if analogue, use a 6 ½ digit whatever
if digital, a freq counter
- then for either just run the math
what am I missing here ?
|
I've only skimmed the more technical documentation on this, (link pHaestus?) but my impression was that the MAX6655 puts two different currents through the diode and measures the voltage drop.
Two different currents need to be applied because the characteristic curve varies substantially from diode to diode, but in a well understood manner. Testing at two different currents allows the IC to compensate for the diode variation by figuring out 'which diode' it is dealing with.
So there are sequencing and processing involved. (I don't know to what extent the processing is digital or analog.)
In terms of you (Bill) using the diode of a
particular CPU to measure temperature, it is trivial. You need a stable current source between 1 and 10 mA, and a good voltmeter. (The current source should have a compliance voltage of 1V or more.)
Just hook up a 4 wire connection to the diode, hook the current source to one pair. (In the right polarity.) Hook the voltmeter to the other pair of wires. Then experimentally determine the voltage vs temperature characteristics of the setup. Your results will have curvature that the MAX6655 does a fairly good job of eliminating, but if you figure out an equation that matches your test results well, you'll have 'plenty' of accuracy and resolution.
When I said 'expensive' I was thinking in terms of something user friendly for those of us without 6+ digit voltmeters at home.