Cryogenic Systems

Cryogenic Wiring Detail

Q: Why is stainless steel coax preferred over copper for cryogenic wiring runs?

Stainless steel (typically 304 or BeCu-jacketed) has a thermal conductivity roughly 25 times lower than copper, so a thin-wall stainless coax such as 0.085 inch (UT-085-SS-SS) conducts far less heat between stages. A 30 cm run of UT-085 stainless coax carries on the order of a few microwatts to the millikelvin stage versus several milliwatts for the same geometry in copper. The penalty is higher RF attenuation, around 5 to 10 dB/m at 10 GHz, but because cryogenic insertion loss adds little to a noise budget on the input lines and the signal is already amplified on the output lines, the heat-load savings dominate the design choice. Where loss must be minimized on the output, a stainless-to-superconducting NbTi transition is used below the first amplifier.

Q: How much heat does a single phosphor-bronze DC wire carry to the cold stage?

Conducted heat is found by integrating the thermal conductivity integral between the two stage temperatures: Q = (A/L) times the integral of k(T) dT. For a 36 AWG phosphor-bronze wire (cross-section about 0.013 mm squared) running 20 cm from a 4 K plate to a 100 mK plate, the conductivity integral is roughly 30 W/m over that range, giving Q on the order of 2 microwatts per wire. A 24-wire loom therefore loads the mixing chamber with tens of microwatts, which is why DC looms are heat-sunk hard at every intermediate stage and why low-conductivity alloys (phosphor bronze, manganin, NbTi) replace pure copper for the inter-stage segments.

Q: At what points must a coax or wire be thermally anchored inside a cryostat?

Every line should be heat-sunk to each temperature stage it passes: the 50 K and 4 K plates of a pulse-tube cryocooler, then the still (around 0.7 K), the cold plate (around 0.1 K), and finally the mixing chamber in a dilution refrigerator. Coax is anchored with bobbin-wound or clamped attenuators and copper bobbins that grip the outer conductor; DC wires are varnished or epoxied onto copper heat-sink bobbins with a length of wire wrapped to give adequate thermal contact length. Skipping a stage lets the line act as a thermal short, dumping heat from the warm end directly onto the coldest plate and degrading base temperature.

/kry-uh-JEN-ik WY-ring DEE-tayl/

The material, routing, and heat-sinking specification that governs how DC bias and RF signal lines run between the temperature stages of a cryostat. The central problem is that every wire is also a thermal short: it carries unwanted heat from warmer stages down to the coldest plate. The "detail" captures the engineering choices that resolve this, including low-conductivity alloys such as phosphor bronze and stainless steel coax, deliberate thermal anchoring at each stage, and the use of inline cryogenic attenuators as heat-sink points on the input lines. A typical dilution-refrigerator microwave channel uses stainless coax from 300 K down to 4 K, then a superconducting NbTi cable on the cold output to recover loss with near-zero added thermal conduction.

Category: Cryogenic Systems

Typical coax: UT-085 SS, 0.085 in

DC heat/wire: ≈ 1 to 5 μW

Managing Heat Load on the Wiring Tree

Inside a multistage cryostat the wiring is treated as a thermal-conduction network before it is treated as an electrical one. A pulse-tube cryocooler or dilution refrigerator offers very limited cooling power at its coldest stages: a typical dry dilution unit delivers only 10 to 20 μW at 20 mK and perhaps a few hundred microwatts at 100 mK. Against that budget, a single uninterrupted copper coax running from room temperature to the mixing chamber would dump tens of milliwatts onto the coldest plate and prevent it from ever reaching base temperature. Cryogenic wiring detail is the set of countermeasures that keep conducted heat within budget while still delivering usable signals.

The first lever is material. Conducted heat scales with the integral of thermal conductivity between the two endpoint temperatures, so designers choose alloys with low and weakly temperature-dependent conductivity for the inter-stage segments. Phosphor bronze, manganin, and stainless steel replace pure copper, and RF lines use thin-wall stainless coaxial cable rather than copper. The second lever is geometry: longer, thinner conductors raise thermal resistance, which is why looms are run with generous slack between stages rather than straight, taut paths.

The third and most important lever is thermal anchoring. Each line is clamped or varnished to a copper heat-sink bobbin bolted to every stage plate so that the heat it carries is intercepted by the stronger cooling power of the warmer stages instead of reaching the coldest one. On the input microwave lines, fixed cryogenic attenuators serve double duty: they reduce room-temperature thermal noise photons traveling toward the device and provide a robust heat-sink point because the attenuator body bolts directly to the plate.

Conducted Heat-Load Estimation

Conducted heat through one conductor segment:
Q = (A / L) × ∫_{T_cold}^T_hot k(T) dT [W]

Where: A = conductor cross-section (m²), L = segment length (m), k(T) = thermal conductivity (W/m·K), and the integral is the thermal-conductivity integral between stage temperatures.

Worked example (36 AWG phosphor bronze, 4 K → 0.1 K):
A ≈ 1.3 × 10⁻⁸ m², L = 0.20 m, ∫k dT ≈ 30 W/m
Q ≈ (1.3 × 10⁻⁸ / 0.20) × 30 ≈ 2 μW per wire

A 24-wire DC loom therefore adds roughly 24 × 2 ≈ 48 μW if it bypassed a stage, which is why each wire is heat-sunk at every intermediate plate.

Wiring Material and Stage Selection

Line type	Material	Conductivity vs. Cu	RF loss (10 GHz)	Use between stages
Input microwave coax	UT-085 stainless (SS/SS)	≈ 1/25	~6 to 10 dB/m	300 K → 4 K
Output microwave coax	NbTi superconducting	≈ 0 (below T_c)	< 1 dB/m	4 K → mixing chamber
DC bias loom	Phosphor bronze, 36 AWG	≈ 1/30	n/a	50 K → cold plate
Twisted-pair (analog)	Manganin	≈ 1/22	n/a	4 K → still
Heat-sink bobbin	OFHC copper, gold-plated	1 (high, intended)	n/a	anchor at every plate

Common Questions

Frequently Asked Questions

Why is stainless steel coax preferred over copper for cryogenic wiring runs?

Stainless steel coax (UT-085 SS/SS) has roughly 1/25 the thermal conductivity of copper, so a 30 cm run conducts only a few microwatts to the cold stage versus several milliwatts for copper. The cost is higher RF attenuation, around 6 to 10 dB/m at 10 GHz, but on input lines that loss is harmless and on output lines a superconducting NbTi segment recovers it. The heat-load saving dominates the choice.

How much heat does a single phosphor-bronze DC wire carry to the cold stage?

Using Q = (A/L) × ∫k dT, a 36 AWG phosphor-bronze wire (A ≈ 1.3 × 10⁻⁸ m²) running 20 cm from 4 K to 0.1 K with ∫k dT ≈ 30 W/m carries about 2 μW. A 24-wire loom adds tens of microwatts, which is why each wire is heat-sunk at every intermediate plate and low-conductivity alloys replace copper.

At what points must a coax or wire be thermally anchored inside a cryostat?

Anchor every line at each stage it crosses: the 50 K and 4 K plates, the still (~0.7 K), the cold plate (~0.1 K), and the mixing chamber. Coax is heat-sunk with clamped attenuators and copper bobbins gripping the outer conductor; DC wires are varnished onto gold-plated OFHC bobbins. Skipping a stage turns the line into a thermal short that dumps heat onto the coldest plate and raises base temperature.

Related Terms

← Cryogenic Wiring Cryogenic Wiring Harness →