From: pcdhSpamMeSenseless@electrooptical.net   
      
   Hi, All,   
      
   In my other thread on the power law circuit, I mentioned the thermal   
   Faraday shield, which may be of interest. (Long-but-worthwhile post warning)   
      
   Simon and I filed a patent last month on a scheme to improve temperature   
   control by a lot.   
      
   *Temperature Control is Slow*   
   Loops controlling the temperatures of macroscopic objects are really   
   slow.  The slew rate is slow because the available heating or cooling   
   power can't move the thermal mass very fast.  More fundamentally, the   
   bandwidth is limited because thermal diffusion is exponentially   
   slow--asymptotically you get another radian of phase shift for every 1/e   
   worth of rolloff.  That makes the usual speedup tricks useless.   
      
   *   
   However, it's possible to eliminate that delay by combining the heater   
   and temperature sensor in a single metal element, such as a bit of   
   copper flex circuit.  This has been done N times before, but apparently   
   nobody noticed one key fact: If you measure the temperature using the   
   heater drive current, *there's no diffusion delay at all*.   
      
   This first came up when I was doing waveguide antenna-coupled Ni-NiO-Ni   
   tunnel junction infrared detectors at IBM, twenty-odd years ago.  Unlike   
   photodiodes, low-barrier TJs work by actually rectifying light, so these   
   were basically crystal radios running at 1.6 um. The devices were about   
   a micron across, made by directional evaporation of gold over nickel,   
   with a short Ni-NiO-Ni junction at the vertex.  The TJ formed a   
   plasmonic traveling wave structure, so that the ~30 fs RC time constant   
   of the Ni-NiO-Ni system didn't trash the response at 190 THz (1.6 um).   
      
   It happened by accident during testing.  This plot   
    is what a reasonably   
   decent device produced in response to a 30-ps pulse at 2.4 um.  (I had   
   this very swoopy tunable laser/optical parametric oscillator system back   
   then.)   
      
   This second plot  shows what   
   happened when one of my tunnel junctions shorted out during testing.   I   
   had about 100 mV of DC bias on it, so even at 3000 ppm/K, a 2.5 mV step   
   is big--a good 8 degrees C.  (In this plot you have to mentally subtract   
   the baseline.)  The bolometric response looked like a step function on   
   this scale, because it took about 5 us to cool back down, even with 3D   
   heat conduction and 1-um size.   
      
   Since the DC current path was about the same as the AC, the   
   near-instantaneous heating of the device produced a near-instantaneous   
   RTD response: about 40 picoseconds.   
      
   Applying this idea to normal life, in principle your temperature   
   controller can have any bandwidth you want.  Of course the slowness of   
   thermal diffusion means that at sufficiently high frequency the   
   temperature of the RTD decouples from the rest of the world.  However,   
   if you tile some surface with these things, you can effectively make a   
   thermal version of a Faraday shield--the huge control bandwidth gives   
   you arbitrarily good rejection of thermal forcing, with no bulky   
   insulation, stirred fluid baths, or big thermal masses.   
      
   The decoupling region actually has some interesting features--as the   
   frequency goes up, the amount of material you have to heat goes down, so   
   there's a region where the phase shift is 45 degrees instead of the 90   
   degrees you get in the low frequency (thermal mass) limit.  (This is   
   discussed in Section 20.3 of my third edition,   
   .)   
   The decoupling also means that in principle the loop bandwidth and   
   compensation don't need to depend on what the element is stuck onto--you   
   get the same huge forcing rejection regardless.  You don't even have to   
   worry about windup, despite the slew rate being slow for the bandwidth.   
      
   There are a number of control schemes for this, of which my favorite is   
   analog PWM.  The heater is in a resistive bridge with a shunt resistor   
   and a reference divider.  It gets turned on for a microsecond or so at   
   the beginning by a strobe pulse and an RS-flipflop controlling an NMOS   
   switch, with the .  A low-noise amplifier (ADA4899-ish) driving a   
   comparator resets the flipflop when the instantaneous temperature error   
   crosses zero.  (The FF is a NAND type, so the heater is turned on if   
   both SET and RESET are active.)  It's cool to watch the duty cycle   
   change instantly if you touch the element, and of course the effective   
   loop bandwidth is huge--a short transient gets nulled out in the very   
   next clock cycle.   
      
   The Class-H thing I talked about in the other thread is for things like   
   DWDM lasers and OCXOs, where you don't want a lot of EMI right in the   
   sensitive region.  (At low power, you can just use an analog loop with a   
   fixed supply.)   
      
   There are a whole lot of things you can do with this general scheme,   
   from improved thermolelectric coolers to such things as a battery   
   calorimeter made of metallized mylar like a chip bag.   
      
   Fun stuff--suggestions for applications and further enhancements welcome!   
      
   Cheers   
      
   Phil Hobbs   
      
   --   
   Dr Philip C D Hobbs   
   Principal Consultant   
   ElectroOptical Innovations LLC / Hobbs ElectroOptics   
   Optics, Electro-optics, Photonics, Analog Electronics   
   Briarcliff Manor NY 10510   
      
   http://electrooptical.net   
   http://hobbs-eo.com   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)