[sdiy] Tempco adjuster idea

[Magnus Danielson]

From: Ian Fritz <ijfritz at earthlink.net>
Subject: Re: [sdiy] Tempco adjuster idea
Date: Sun, 27 Apr 2003 11:49:27 -0600

> Hi Magnus --
> 
> At 09:30 AM 4/27/2003, Magnus Danielson wrote:
> > >
> > > Actually, the kT comes from classical statistical mechanics
> > > (thermodynamics), but thanks for the support!  :-)
> >
> >Yes, it does... and to show why they invented the quantum mechanics, since 
> >it prooved
> >to be part of quantum mechanic behaviour, where statistical mechanics is 
> >mearly an
> >extention to that field (even if the areas where originally invested the 
> >other way
> >around).
> 
> Well, sure, but I thought you were saying that the semiconductor energy 
> levels were coming into play.  Those go into the I_es prefactor.

There are _several_ energy levels in play. But kT forms the energy level for which
the qVbe energy level is referenced to form the statistical properties. Check out the
units ;O) Checks out very well with statistical mechanics which also has the e^(E/Er)
form.

> >That takes carefull measurements. I have two good DMMs, but no means to 
> >measure
> >temperature very accurately. It seems nobody really seems to give out the 
> >details on
> >that. Any proposals?
> >
> >Ease of calibration is my concern. (Ice-water and boiling water in my lab 
> >doesn't
> >*really* feel like a good idea.)
> 
> I just use an LM335, which directly reads absolute T at 10 mV/K.  You can 
> correct the readout with a single resistor determined from a single T 
> calibration point if you want, but so far I haven't bothered.

Seems reasnoble.

> For the next level of accuracy and precision you can use a platinum (Pt) 
> sensor -- an RTD element -- in a Wheatstone bridge circuit.  You can buy 
> these elements precalibrated.

Hmm... sounds like another option to investigate. If I recall things correctly you
should have a reference platinum sensor, but I guess that depends on how you rig it.
Maybe a 4-point resistance measurement is a better option than the Wheatstone bridge.

> For a calibration point you can use the melting point of ice if you are 
> careful.  The ice from your freezer will be below the freezing point, so 
> you need to crush it up real fine and make a slurry with some water inside 
> a thermos bottle.  Then immerse your sensor in oil inside a test tube, 
> making sure it doesn't touch the glass, then put that assembly into the 
> ice-water slurry.  This is a standard lab procedure for making a 
> thermocouple reference junction.   To be super-fussy, use distilled or 
> deionized water for both the water and the ice.

Sounds like an interesting exercise, but what I really wanted to avoid... but maybe I
should see it as my excust to get back to chemestry... ;O)

> Boiling water is not recommended for calibration, for several reasons 
> including lack of homogeneity, dependence on atmospheric pressure and 
> dependence on impurities.  It's OK for rough estimates, though.

Ah, right. I have a real Hg barometer thought. Used to be used in my fathers chemestry
lab.

> The biggest problem with measuring drift of a VCO is keeping the 
> temperature uniform across the circuit board.  Gradients and drafts just 
> kill you.

I've noticed. I had problem with "industry standard" oscillators (they are pure
crap!). I had to close my window (it was a hot day in the lab) and continue the
measurement. Draft counts! Crystal oscillators are temperature senstive... 

> I put my boards in a plastic container with crushed-up newspaper 
> to prevent drafts, and I try to heat the container uniformly.  But it is 
> not easy to measure drifts below about 200 ppm/K.

Seems like a good idea. Double-oven is what is used for higher precission stuff.
What you do there is that you have a metal case (good thermical conductor) followed
by a good isolator, with a metal case again and then good isolator before hitting the
actual device. For precission devices you also care about thermical conduction of
the support.

> >Ah! Actually, I looked at Jim's design and considered an alternative 
> >design. Jim used
> >up a complete OTA for the inverse properties, where as I worked on how to 
> >do things
> >properly with the use of linearizing diodes. I put it away, but I felt 
> >quite confident
> >that reductions of Jim's curcuit could be done.
> 
> Jim tells me he looked at the linearity of several configurations and 
> picked the best one.  I think linearity is the biggest issue with the 
> active compensation scheme.  In the double OTA scheme Jim uses, the 
> nonlinearities cancel because the two inputs are at the same voltage.

Well, I did actually look on how the linearity diodes *really* works, and I have yeat
to see them used correctly. The standard setup is an approximation. That's why it's
not very linear. I see nothing wrong with going back to the linearized diode plan.

> >The whole trick where to do a "bandgap" temperature reference and multiply the
> >incomming CV with that reference, that was the idea. This multiplication 
> >naturally
> >needs to be very linear... and you end up having some problems in another 
> >corner.
> >However, such things can be solved. Once you know that general plan, it's the
> >execusion of the plan which should form the next layer of confusion.
> 
> What I meant about not understanding Jim's circuit is that I can't get the 
> numbers to come out right.  The circuit analysis seems straightforward, but 
> I get the wrong scale factor.

Which way? What numbers do you get?

> >The comparator seems to be forgotten all the time BTW. I've never seen an 
> >analys of
> >its sensitivity from external sources (temperature included) into the 
> >frequency of the
> >oscillator.
> 
> Agreed -- this should be looked at more.  Any ideas on what to measure?

A good indication of something being wrong there would be to measure the peak values
or maybe average values. Then your attention should be drawn to the comparator.
Troubles is, you must make very accurate measurements, since the flaw in amplitude is
linear to the flaw in frequency.

> >So you are saying that I should not heat things to 1500 K unless I take 
> >out the
> >FET-opamps? ;O)
> 
> Or just use them as fuses.  :-)

I don't take heated argument like that so well ;O)

A fuse without forced cooling is a fuse which has not been able to show it's
potential!

Cheers,
Magnus