18
\$\begingroup\$

I am reading this paper : Upgrade of a low-temperature scanning tunneling microscope for electron-spin resonance and found that for thermalization, they use 0 dB attenuators.

Page 6: Having the finite cooling budget of our cryostat in mind, and noting that the heat transfer of the RF line is dominated by the micron thick Ag coating of the center conductor, we thermalized all SR cables using 0dB attenuators

I understand attenuators reduce signal intensity (power), but what does it mean to have a 0 dB attenuator?

What is the general purpose of 0 dB attenuators and in the paper is it just being used for thermal coupling?

\$\endgroup\$
  • 1
    \$\begingroup\$ Have a look here. This should give you a good idea of what was going on, and why. The 0 dB connector (which isn't exactly 0, but might be as much as half a dB) is a vital part between their semi-rigid coax segments between systems of different temps. In reading, they really had a hard time setting things up and maintaining them, too. They were working hard. A datasheet for one of their connectors is XMA Corp. 2782. \$\endgroup\$ – jonk Nov 6 '18 at 20:18
  • \$\begingroup\$ @DwayneReid Thanks for your comment, I removed my answer and will check in it (and read the other answers/comments). \$\endgroup\$ – Michel Keijzers Nov 6 '18 at 20:23
  • \$\begingroup\$ Related \$\endgroup\$ – user3490 Nov 7 '18 at 13:30
  • \$\begingroup\$ Thanks for the question! It was interesting to learn about experimentation using electronics in cryogenic chambers. \$\endgroup\$ – user203591 Nov 8 '18 at 1:46
9
\$\begingroup\$

JRE's answer covers why 0dB attenuators exist in general, in this specific case it sounds like they're using them to make good thermal contact between the cable's inner conductor and the thermal isolation stage as you can see in the picture on page 20, and thus to the cryogenic reservoir which is cooling the thermal isolation stage. Basically, you want the cable to be at the same low temperature as the stuff down in your cryogenic can, rather than conducting all the heat from the connection up top at room temperature. An added problem is that cables have anisotropic heat conduction: they conduct heat better along their axis than they do radially, because the insulator that keeps the centre conductor electrically isolated is also a good thermal insulator (or at least a lot better than the metal of the inner conductor). Attenuators have better thermal conduction between the outside and the centre, and so they do a better job of thermally connecting the inner conductor of the cable to the thermal isolation stage.

Judging from the mounting it also provides some mechanical stability, the attenuator is held in one place while the cables attached to either end will move around a bit due to thermal contraction as you go from room temperature down to 4K or lower.

Cryogenics is hard, and heat transfer from the rest of the horrible 300K world is your greatest enemy.

Generic cryogenic RF apparatus sketch


Page 20

enter image description here

\$\endgroup\$
  • \$\begingroup\$ Thank you for the response. So, in this paper, they use 0dB attenuators in order to have better thermal conduction? It sounds like 0dB attenuators are going to conduct a lot of heat from air to the mK stage? I thought you generally want to have good thermal isolation between stages. Is the purpose of the 0dB attenuators to help the inner conductor of the RF line thermalize but the rest is thermally isolated between different temperature stages? \$\endgroup\$ – Blackwidow Nov 7 '18 at 14:06
  • 1
    \$\begingroup\$ Right, the idea is that the attenuators have better thermal conduction between the thermal isolation stage and the cable's inner conductor (Ofer's answer made a good point I didn't think of, I'll sketch something up to try and illustrate). They don't really affect the thermal conduction from the outside world, or between thermal isolation stages. \$\endgroup\$ – llama Nov 7 '18 at 23:05
  • \$\begingroup\$ Another way you can think of it: the attenuator links the upper (higher T) segment of the cable to the thermal isolation stage (TIS), so that the heat which it conducts is dumped into the TIS instead of continuing on down the cable into the really sensitive area. TISes are usually thermally linked to a reservoir of liquid helium or nitrogen, so they can sink a relatively large amount of heat load, but at a higher temperature than the refrigerator-cooled region. \$\endgroup\$ – llama Nov 7 '18 at 23:08
  • \$\begingroup\$ This line seems to contradict the rest of the post, some editing may improve the quality of the answer. "in this specific case it sounds like they're using them to damp changes in temperature from RF pulses in lossy cables" \$\endgroup\$ – user203591 Nov 8 '18 at 1:44
26
\$\begingroup\$

A 0dB attenuator is an attenuator that is electrically and mechanically just like other attenuators in that product line.

You can use them in place of another attenuator when doing tests. It acts just like the other attenuators (in terms of insertion loss and frequency range) and fits in exactly the same space mechanically.

A 0dB attenuator meets all the specifications of the others in the same product line, so swapping it in keeps all other performance characteristics of your setup the same.

Might not matter if you are tuning a CB radio, might be critical in high performance, cutting edge stuff.

The point of using one in the referred paper seems to be to reduce the amount of heat seeping into a very sensitive piece of equipment. Heat causes noise in the extremely low temperature circuits the paper describes.

The 0dB attenuator seems to be used because it passes the RF signal (almost) as well as the cable, but doesn't pass heat as easily as the cable would.

\$\endgroup\$
  • \$\begingroup\$ thank you for the answer. Your answer sounds right and given the fact that they point out their 'finite cooling budget,' your view makes sense. I was just confused about the sentence: "We thermalized all SR cables using 0dB attenuators." To thermalize all cables means to make them have thermal equilibrium.... \$\endgroup\$ – Blackwidow Nov 6 '18 at 20:43
  • 3
    \$\begingroup\$ @JRE It seems that a major objective when passing cables into cryogenic chambers is to ensure that the cable is fully cooled to the chamber temperature so that it does not directly warm the device in question through conduction. With coaxial cable, cooling the inner conductor seems to be of particular interest/difficulty. Is there some mechanical property of an attenuator that might help cool the center conductor if it were thermally bonded to a very cold piece of metal or floating in cryogenic liquid? \$\endgroup\$ – user203591 Nov 7 '18 at 0:45
  • \$\begingroup\$ @JRE looking at the responses posted below, it sounds like you want the inner conductor of the RF line to thermalize (inner conductor of the RF line is cooled rather close to the mK stage) while maintaining good thermal isolation elsewhere? \$\endgroup\$ – Blackwidow Nov 7 '18 at 14:09
3
\$\begingroup\$

A 0-dB attenuator helps with physically thermalizing the center conductor of the cable. The attenuator substrate is usually a better thermal conductor than the cable’s Teflon dielectric. They don’t, however, help with noise coming down from 300K electronics. For that you would want attenuation equal to or larger than the ratio of room temperature to the temperature of the attenuator. For example, you would want a 20 dB attenuator to reduce 300K noise to 4K level.

\$\endgroup\$
  • 4
    \$\begingroup\$ Could someone explain just exactly what "thermalizing" is? It obviously has to do with temperature, but the referred paper and everything I've been able to find on the subject just assumes that everyone knows what thermalizing means. \$\endgroup\$ – JRE Nov 7 '18 at 9:40
  • \$\begingroup\$ I'm not sure if I've ever heard a concrete definition, but vaguely that the thermal connection between them is good enough that they'll always be at the same temperature. Maybe something more precise along the lines of at thermal equilibrium (once you've filled the thing up with LHe and waited for a while), the temperature differential between them will be < x (milli)kelvin. \$\endgroup\$ – llama Nov 7 '18 at 22:57
  • \$\begingroup\$ similar to "winterizing"? \$\endgroup\$ – Misunderstood Nov 11 '18 at 20:10

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.