Importance of matching in expo convertors?!

[Rene Schmitz]

Hi Ben!

At 08:22 31.08.99 -0600, you wrote:
>Being new to this mailing list, I was wondering if PTAT (Proportional to 
>absolute temperature) current sources were ever used to compensate for the
>temperature dependent gain of exponential voltage to current converters.
>(PTAT sources are ubiqutous in IC design to amplifier gains stable over 
>temperature and to create voltage references).
>
>I can see how just using a differential pair will create a temperature
>stable reference
>point, but the gain will still drift over temperature. Once way to prevent
>this from 
>happening is to put a heater on the differential pair.  Another way would be
>to 
>use a PTAT current source....

What we need is a multiplier/divider then. This is the technique used in
the CEM chips.

>PTAT current sources are easy to build on chip, but I don't have any
>experiance
>doing this on the board level.  Somthing like the LM134 might work...

My "temperature compensation amplifier" started from the same idea, a
voltage/current inversely proportional to absolute temperature and multiply
the input sum with that voltage. But I realized that I would have to
overcompensate for the multipliers gain-tempco as well, and this is
difficult, because the multiplier (e.g. an OTA ) would not be on the same
temperature than the expo pair. I realized that the OTA or a diffamp has
the same temperature dependancy than the expo, thus I put it in the
feedback loop of an opamp, to turn the dependancy into its reciprocal. So
the drift can cancel. To ensure that the thing works, one has to use a four
transistor array, I build a diffamp from two and the expo convertor from
the other two transistors. The main drawback here is that you get a little
nonlinearity from the diffamp, but it is still usefull. And at that low
signal levels the opamps offset drift becomes more of a problem. (One can
tweak the circuit still a little towards less THD. I did calculate this but
never tried it.) You can see the schematics on my website. The whole "TCA"
is a inversely proportional to abs. temperature gain amp. But all
transistors being on the same die, one might say this is also an on chip
technique.

Bye
 René




error:asciiartisnotfully  \ uzs159 at uni-bonn.de
supportedwithyourcurrent   \ http://www.uni-bonn.de/~uzs159
displaydriversettings...    \ http://members.xoom.com/Rene_Schmitz