I want to know what kind of improvemend electronic circuits get at cryogenic temperatures.I looked at online noise calculator and it shows that if I go from room temperature to 1 kelvin the noise drops by 3db,that is lot less that I expected.

So if I have galium arsenide transistor at room temperature and then I cool it with liquid helium to some very low number like 5 kelvin,is 3db really all i get? Does the noise drop proportionaly across the whole spectrum from DC to the maximum freqency of the transistor,or does the noise drop more in specific part of spectrum,for example if the temperature have more effect on 1/f noise in the lower freqencies.

It seems to me the mild drop in noise doesnt justify the added expense of using cryogenic electronic circuits,can you name me the benefits of running amplifiers at low temperatures?

  • \$\begingroup\$ Yeah. Cool it with liquid helium, this a very usual way the electronic stack exchange members cool their devices. \$\endgroup\$ May 16 '17 at 16:51
  • \$\begingroup\$ Electronic devices are likely to decompose at "cryogenic" temperatures, as the solder tends to change it's properties. Batteries will die too. Semiconductors turn insulators as there is much less thermal excitement of the free charge carriers. Resistors turn superconductors. So yeah, pure improvement. \$\endgroup\$
    – Eugene Sh.
    May 16 '17 at 16:54
  • \$\begingroup\$ @MarkoBuršič Liquid nitrogen (77K) is a lot cheaper and more practical than helium which is a precious finite resource on earth. Any helium that is released from natural gas wells into the atmosphere disappears into space. \$\endgroup\$ May 16 '17 at 20:32

Resistors drop in noise proportional to the square root of temperature, so at 4.2K a resistor will have about 1/8 the noise as at 300K.

Bipolar parts usually stop working entirely.

Some GaAsFets apparently work okay, MOSFETs can work okay.

The noise reduction will depend on the source of the noise. If it's white noise based on resistance it will decrease with sqrt(T). If it's shot noise it will depend on the barrier height. Flicker noise (1/f noise) may be a bit less predictable.

Is it worth the cost? If you can't accomplish the goals any other way, then it may well be. It's possible to achieve measurements that are impossible any other way using superconducting quantum interference devices.

  • \$\begingroup\$ 1/17? But 3db decrease is like 2/3,thats huge difference! I dont understand where does this difference in our numbers come from,I saw on multiple online noise calculators that if I go from room temperature to 1 kelvin I gain no more than 3db \$\endgroup\$ May 16 '17 at 18:13
  • \$\begingroup\$ @wavscientist 1/8 in voltage (1/71 in power) -- I must have rushed that calculation. What calculators are you referring to? \$\endgroup\$ May 16 '17 at 20:26
  • \$\begingroup\$ Spehro Pefhany that is still big difference. rfcafe.com/references/calculators/… \$\endgroup\$ May 16 '17 at 20:43
  • \$\begingroup\$ @wavscientist At 0K noise temperature of the amplifier, the amplifier will add nothing to the (say) 50 ohm resistor at the input (assumed to be at 300K). At 300K it adds 3dB (doubles the power). But if you chill that 50 ohm resistor down to 4.2K resistor things change. \$\endgroup\$ May 16 '17 at 21:29
  • \$\begingroup\$ I still dont understand how you come up with 1/8 voltage and 1/71 power,these numbers are far greater than 3db.Your numbers are bigger becose at lower temperature the resistance of transistor in ON state decreases so its the combination of lower thermal noise and lower resistance noise that give your "1/8" number? \$\endgroup\$ May 17 '17 at 0:05

NASA uses 8GHz MASERS at focal point of their 3 dishes, 70 meters each, at Madrid/GoldStone/Canberra for downlinks at 20 degrees Kelvin. The actual temperature of the MASER is 14 degrees Kelvin, but some noise contributors (metal pieces) raise the effective noise temperature to 20K.

Now............how much "electronics" is included in that MASER amplifier?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.