Bench Talk

Researchers Develop an Amplifier That Only Wakes When the Qubits Call

(Source: Влада Яковенко/stock.adobe.com; generated with AI)

Many quantum computing headlines focus on one thing: the qubit. They get all the attention in quantum computing because they have enough noteworthy features to be impressive, and are simple enough to explain. Furthermore, the ever-increasing number of qubits being packed into quantum solutions makes for an eye-opening story of seemingly limitless progress.

But most headlines don't talk about the hardware, unless it fails, of course. In quantum systems, hardware components matter even more due to the razor-thin margins and cryogenic environment.

While quantum computers already exist, making them bigger and better isn't easy. The physics is here, but the real magic lies in how to engineer this technology to scale. Behind every functioning qubit is a stack of supporting hardware: amplifiers, filters, switches, cables, cryogenic enclosures, and control electronics.

As researchers push toward machines with hundreds or thousands of qubits, the engineering challenge is centered around wiring, power, cooling, and performance. Engineers need to build smarter components that adapt to quantum's unique constraints, especially when it comes to power, timing, and cryogenics. This blog looks at an ongoing effort to meet this challenge.

Rethinking the Amplifier

At Chalmers University of Technology, PhD researcher Yin Zeng is rethinking one of the most underappreciated pieces of the quantum puzzle: the microwave amplifier (Figure 1).^[1]

Figure 1: Yin Zeng, one of the study’s authors, mounts the newly developed amplifier into a quantum computer’s cryostat at Chalmers. (Source: Chalmers University of Technology, Yin Zeng/Maurizio Toselli)

Zeng's research starts with a simple question: can we reduce power dissipation by activating the amplifier only during qubit readout?

The idea came from a very real problem inside the dilution refrigerator. Amplifiers, unlike most other components in the system, are active, which means they generate heat, and a lot of it.

"Our amplifier takes about 99 percent of the cooling capability," Zeng explained.

That thermal load doesn't just tax the cooling system. It also risks interfering with the qubits themselves. "Qubits are very sensitive," he said. "They pick up all the noise and influence of the environment."

Zeng thought about this differently, reconsidering if the amplifier needed to stay on all the time? In a typical quantum computing cycle, the readout window—the time when a qubit's signal actually needs to be amplified—is just a fraction of the total operation. If the amplifier could be turned off during idle periods, that could dramatically reduce the thermal load without sacrificing performance. This is where a pulsed high-electron-mobility transistor (HEMT) low-noise amplifier (LNA) proved beneficial to their research (Figure 2).

Figure 2: Conceptual illustration of a pulsed HEMT LNA operating inside a dilution refrigerator. The amplifier is activated only during the short readout window, reducing heat dissipation without compromising performance. (Source: Chalmers University of Technology, Yin Zeng)

Of course, turning an amplifier on only when needed is easier said than done, especially in a cryogenic quantum system where timing, noise, and readiness must align within nanoseconds.

The team had to determine how long it takes for the amplifier to be fully ready once it's switched on. Without that information, timing amplifier activation with qubit readout would be imprecise.

"We used a very high-speed oscilloscope to build a setup that can characterize the noise performance or noise transient in time domain with nanosecond resolution," he explained (Figure 3). The goal was to determine when the amplifier's noise returned to baseline, which indicated it was stable and usable for quantum signal processing.

Figure 3: Experimental setup used to measure noise performance transients of the cryogenic amplifier. This setup enabled the team to determine, with nanosecond accuracy, when the amplifier becomes stable and ready for use. (Source: Chalmers University of Technology, Yin Zeng)

What they found was that a typical square pulse left the amplifier too slow to respond. It could take up to 200 nanoseconds to return to optimal performance. In a system where the entire readout window might only last 300 nanoseconds, that delay made the smart on/off idea impractical.

To solve this, Zeng and team sent out a pilot pulse instead of a standard square wave to turn the amplifier on. They designed a custom waveform that could bring it to readiness faster without compromising noise performance.

With the new pilot pulse, the amplifier's wake-up time dropped from 200 nanoseconds to just 35 nanoseconds—fast enough that the amplifier can be reliably activated only during readout, improving thermal efficiency without compromising signal integrity.

Avoiding the Usual Trade-Offs

One of the most exciting aspects of the work is that it avoids the compromises typically associated with reducing power consumption. In most amplifier designs, especially those operating at cryogenic temperatures, cutting power usually means losing something, such as gain, bandwidth, or noise performance. In this case, the amplifier itself wasn't fully redesigned. The improvement came from how it's operated, not how it's built.

This means the amplifier still performs at the levels required by quantum systems, but without burning power the entire time it's idle.

Alleviating System-Level Constraints

This new operational approach has implications beyond the amplifier itself. Heat in a dilution refrigerator could affect the entire system budget, including how many qubits can be supported, how isolators are deployed, and how big the fridge needs to be.

"If we have like 10 times lower power dissipation, then that just released a lot of space design for the heat dissipation," Zeng explained.

Making amplifiers more efficient can ease constraints across the board. In some cases, Zeng said it may even reduce the number of isolators needed to block back-propagated noise from interfering with qubits.

"There was a paper actually talking about this … They found that the amplifier dissipation power influences qubits, so they have to add two more isolators," he said. "Maybe we need only one."

Engineering the Quantum Stack, One Layer at a Time

For Zeng, this work isn't just about amplifiers. It's about expanding the mindset around how quantum systems are designed, especially as they go from the lab to scalable platforms.

"From one to a thousand qubits is less a physics problem now; it's more of an engineering problem," he said.

According to Zeng, much of the research comes down to optimizing across multiple dimensions. Traditional amplifier design tends to focus on frequency domain performance, gain curves, bandwidth, and noise temperature. Zeng’s research adds a new dimension: time.

"We sometimes need to think outside the box," he said. "I'm happy to say that my research explored the time dimension in this design for circuits."

Conclusion

The Chalmers team challenged the assumption that every component needs to work the way it always has. By exploring timing as an engineering dimension and forcing traditional hardware to behave differently in a quantum environment, they are helping carve a path toward systems that are both functional and scalable.

In a field often dominated by big qubit counts and larger promises, it's the overlooked advances, like smarter amplifiers, cleaner signals, and cooler fridges, that could make the biggest impact.

Zeng and the team have made their published paper open access on IEEE TMTT, so engineers can learn more.^[2]

Sources

[1]  https://www.chalmers.se/en/current/news/mc2-smart-amplifier-enabler-for-more-qubits-in-future-quantum-computers/
[2]  https://ieeexplore.ieee.org/abstract/document/10969553