On September 1, we launched the second generation of our readout analyzers for superconducting and spin qubits - the SHFQA Quantum Analyzer. On November 17, we dedicated an hour-long event to provide you with the latest insights into the instrument, demonstrate key features and answer your questions. If you missed the event, or if you would like to re-watch parts of it - please find the recording here.
In the first part, you learned from our CEO Sadik Hafizovic about Zurich Instruments' quantum strategy and the role of the second generation of our Quantum Computing Control System (QCCS) - in specific of the SHFQA. Paolo Navaretti, who also led through the event, then provided you with an overview of the the SHFQA's main characteristics and key features as well as its main application: multiplexed qubit readout of up to 64 qubits.
Later, we discussed more details of the technical applications, features and also showed first results on real qubits. We interleaved the presentation with demonstrations using a real SHFQA instrument, and specifically:
- Setup and control of the integrated frequency up/down-conversion
- Direct resonator spectroscopy of a 3D cavity at 8.1 GHz,
- Readout of 16 qubits in parallel and with high fidelity,
- Integration of the SHFQA into a large QCCS with 16 instrument while running a synchronized Ramsey sequence on all instruments
- A global, active-reset experiment feedback involving our central controller, the PQSC
In the last part of the event - the Q&A session - we did not have time to answer all questions. In the remainder of this blog post, we therefore address all questions again. They have been sorted by subject for easier navigation.
Enjoy browsing them and - as always - if you have any questions, please get in touch.
SHFQA General Questions
When will the SHFQA be available?
You can order now, shipping of the first devices will be in the first quarter of 2021, so only a few months away.
Can I also buy an instrument with only 1 channel?
Short answer is no, we have a 2-channel and a 4-channel version.
How much will the new device cost compared to the UHFQA?
This depends on the exact configuration that you need for your experiments. Please just contact us and we are happy to configure the system to your needs and provide a price.
I need code for certificate. How can I have it?
Please contact us, and we will be happy to help.
Could you explain: How are you "generating" a quantum signal? Polarized photons, electron spins… or is this a system just for controlling a generated quantum signal?
[See recording for more information] With our quantum analyzers, we are not generating quantum signals itself, but rather determine the state of a quantum bit. For this, we generate a classical microwave pulse the amplitude and phase of which are being modified by the quantum system. This change is then detected by the SHFQA in real-time and with maximal signal-to-noise.
Could I read out my neutral atoms with the SHFQA?
The short answer is no, it was not designed to do that. However, applications and systems exist where an SHFQA could be beneficial to do just that. We are always interested to expand our application fits, so if you want to discuss on how to control or read out your neutral atoms - please contact us.
How do you manage to make multiple Quantum measurements not destroing Q-bits? Probably a bit technical but, could you elaborate how can you "control" a quantum signal without collapsing it or measuring it?
[See recording for more information] Let me answer these questions together. Dispersive readout, a type of quantum non-demolition measurement, exploits that the readout resonator is very far detuned from the qubit itself. This means, that the resonator frequency provides information of the population of the qubit - is it excited or not - while it does not affect the state of the qubit itself. This means, simply spoken, that repeated measurements of the qubit state will provide the same results, i.e. the readout process does not change the free evolution of the qubit state (e.g. it does not excite it). It still projects or collapses the state, but the measurement outcome uncertainty does not increase with repeated measurements.
By this "resonator" do you mean you can actually clone the quantum state of a qubit signal?
No, dispersive readout using a readout resonator neither provides the full information about the qubit state in a single measurement, nor can this readout clone the qubit state quantum mechanically.
SHFQA Technical Questions
How to get rid of IQ mixers? By using direct digital downconversion? Ultrafast ADCs?
As discussed in the first part of this event, we use a double super-heterodyne approach that uses a combination of mixers and filters to provide a broadband and clean analysis band. We are convinced that this is currently the best approach when compared with IQ mixers or direct RF approaches using fast ADCs and higher Nyquist zones.
You sample a signal at 3 GHz with 4 GS/s. What about Nyquist criteria and aliasing?
[See recording for more information] Indeed, we are sampling in the second Nyquist zone. However, this does not result in a lower performance of the digitization and provides the added benefit that the digitized signal can be presented to the user similar as in the signal generation. Given that a Nyquist zone extends over 2 GHz, it covers the SHFQA's analysis band of 1 GHz well and aliasing signals can be efficiently filtered.
May I know the readout latency value of SHFQA?
We are still working on a proper characterization of the readout latency of the instrument, which we consider a key feature that will be continuously improved. The results will be published on the website before February 2021 and are expected to lie below 200 ns.
To optimally readout qutrits, how many signal generation units do I need?
[See recording for more information] To optimally distinguish the three positions of your readout resonator when reading out qutrits, one ideally uses 2 readout tones that maximize distinction between two states. In the SHFQA, this can be done by either programming two waveform memory blocks and separately control them, or by combining the signals into one block and use the remaining one for other manipulations of the readout resonator.
Compared to the UHFQA, how much is the difference in phase noise (or general noise) contribution to qubit control?
[See recording for more information] It is not so easy to compare them directly as for the UHFQA you need to supply your own frequency-conversion electronics, but not on the SHFQA. Typically, the synthesizers in this conversion process are what is limiting the phase-noise performance, not the signals in the base-band. We put a great deal into designing our synthesizers such that they are not the limiting factor when reading out qubits. If you want to know the exact specifications, please contact us and we can provide them to you.
SHFQA and QCCS
Can you explain a bit more what the ZSync is?
[See recording for more information] The ZSync is our own proprietary link between the control instruments and the central controller. Why another 'proprietary' link? The key is, that for quantum error correction other architectures (e.g. PXI) do not offer very low and consistent latency in conjunction with a central node of knowledge. The ZSync is a very simple protocol tailored to offer just that and system synchronization.
How much latency do you have with feedback?
This question depends strongly which feedback path one considers. For the demonstrated - most demanding - global feedback through the central controller, the PQSC, we showed a preliminary feedback time of ~800 ns. However, we are confident that with improvements to the system over the next few weeks, this number can be pushed well below 700 ns, which is the latency of our first generation Quantum Computing Control System (QCCS).
What other instruments can we expect in your second generation of control system?
About 50 people are working on new instruments so you can expect new technology coming out of our pipeline. Unfortunately, we cannot say much more than fueling your imagination further - but one of the main innovations we are currently working on is to get rid of all IQ-mixers in a control system.
When will you have a system that can control 1000 qubits?
[See recording for more information] With the second generation of our Quantum Computing Control System, we will be able to support at least 100 qubits, with 200 well within reach. As of today, providing a control system that controls 1000 qubits does not make sense as there is no hardware in need of that. We are confident, though, that we will have a system in time.
Do you plan to help customers with the challenge of fitting 100s of cables into the dilution fridge?
This is not our primary area of expertise. However, we are fully aware and informed of the upcoming challenges related to that and closely collaborate with relevant partners through our joint research projects, e.g. Bluefors through OpenSuperQ. That is why we - already today - implement and design features that will reduce the number of cables, e.g. through multiplexing capabilities, and we keep up-to-date with new approaches and its corresponding requirements.
Is Mark Zuckerberg trying to spy us with Quantum Computers?
If he is, then he is not using our Quantum Computing Control System - that might be a mistake.
Zurich Instruments and its quantum strategy
Are you participating in government-funded research projects?
[See recording for more information] Yes, we participate. They are not part of our business model, but how we do innovation. The SHFQA is a great example for that as its concept and design were developed in the IARPA-funded QSurf project.
On the chart with the qubit numbers, can you comment on the 100-qubit system data point at the end of next year?
[See recording for more information] We want to make sure that we are never the limiting factor in research. Our leading customers target 100 qubits on that timescale so we make sure that we have an appropriate system that will fit into two racks.
Do you have any advice for a high-school student who wants to pursue a career in quantum computing?
[See recording for more information] We are all not aware of the full potential that quantum computing will bring in the decades to come - and the community will need a lot of intelligent people to figure out and unlock the full potential of quantum computing. Great that as a high-school student you are looking into that. Find a place where quantum computing is taken serious, be sure to find the right mentors, be selective in choosing your place, and the time now is great because everyone is looking for good people.
Do you have any PhD position or any internship?
We are always looking for talented and motivated candidates that work with us, for example through an internship. If you are interested in becoming part of Zurich Instruments, please check our current openings or contact us.
How do you see the future of qubits?
Bright and shining :-)