[sdiy] tempco stuff

[ryan williams]

hi all,

I'm writing a paper for an engineering communications class. class about 
speaking & writing, not communication systems (already had that one). We 
have to write a technical paper on something that interests us. I chose, 
temperature compensation in musical instruments. I will not have time to 
actually build things and take measurments. I would like to, but there 
isn't much chance I can blow off my EE projets for that.

Anyhow, I had some questions. For tempco resistors, often I see people 
use resitive dividers. It seems to me that this can allow pretty good 
compensation even with non 3353ppm resistors (number I calculated as 
ideal assuming 25degC is room temperature).  What I am trying to find a 
good resistor ratio for the divider. I am picking values at the 
endpoints of some temperature range I think the device will work in. I 
chose 77F to 100F (25C to 38C) then solve for a resistor so that the 
ratio of Vin*Divider/Vt is the same for both temperatures. I have not 
yet included the tempco of the normal resistor. I get almost exactly 
24Kohm with a 1K 3500ppm tempco. Any other scaling of the input to get 
1V/octave would be in an opamp before the resitive divider. Is this a 
reasonable method? Any other ways to adjust the tempco using other circuits?

for electronic compensation. Does anyone have any comments on a Vt 
reference that I could use for scaling the input voltage. I have seen 
this design, called the Brokaw Cell (pg 299 of opamp applications 
handbook). And I have thought of another similar circuit using two 
opamps and a feedback loop on each to keep two transistors of a diff. 
pair's collector currents constant. then the base voltage on one is 0V 
and the other is controlled by one of the opamps. The emitters are 
controlled by the other opamp. The output would be the base voltage as 
some constant ln(Ic1/Ic2) multiplied by Vt. I'm wondering about other 
possible references. I have seen Jim Patchell's but do not know enough 
to compare it with other circuits. Anyone know of something written on 
this subject?

I have lot of plots to make before this thing is over. In the end, I 
plan to put it on my web page because it might be helpful for people who 
know <= myself.

-ryan