The manuscript entitled “Fast and sensitive measurements of sub-3 nm particles using condensation particle counters for atmospheric rapid measurements (CPC FARM)” by Darren Cheng and colleagues reports on a new device capable of size-resolving particles with diameters below 3 nm utilizing condensation techniques. The authors use a battery of water-based condensation particle counters based on the design of the MAGIC 250 from Susanne Hering but operate the individual devices at different temperature settings to obtain a range of cut-off diameters. By taking differences of the measured number concentrations in the different channels the size distribution is recovered without the help of pre-size selected differential mobility analyzers. As concentrations are therefore much higher in the detectors the sampling rates allow much faster monitoring of the dynamics during new particle formation compared to an SMPS system suffering from low counts for several reasons.

The manuscript starts with an introduction of current techniques relevant to atmospheric aerosol sizing in the smallest possible size range which very much goes along my own thinking. Followed by a brief description of the general features of the CPC FARM, the instrument’s performance is characterized by calibration studies using state-of-the-art techniques. Extensive analysis of potential uncertainties in the CPC detection is performed and subsequently data inversion procedures are described in detail. The performance of the CPC FARM has been tested with ammonium sulfate particles and ambient air and comparison to an SMPS has been made demonstrating the improvement in detection for particle sizes below 3 nm. Furthermore, data inversion techniques are discussed using synthetic particle size distributions and simulated instrument responses. Overall, the manuscript is of very high quality and clearly fits the scope of Atmospheric Measurement Techniques. Below I’m listing a few mostly minor things that might be considered before final publication.

Section 2.1 describes the CPC FARM in a very general way. I understand it is based on the MAGIC 250 instrument (and the TSI 3789) but still some more details would be helpful. I’d recommend putting a schematic into the supplemental material illustrating the core setup of a single channel. Also, the temperature tuning of the different channels could be addressed in this section. It is described later on in section 2.4 (Pittsburgh Campaign Setup) but I assume the same settings have been used already for the CPC FARM experimental characterization (section 2.2).

One thing that should be further discussed with respect to the performance of different channels in the CPC FARM is the behavior of detection efficiencies. It appears to me that channel 3 systematically shows larger cut-off diameters and reduced slopes of the detection efficiency curves compared to all other channels (figure S1 in the supplemental material). Does this in any way impact the data inversion? It is also interesting from a technical point of view as the same geometry and operating conditions not necessarily gives the exact same output in laminar flow devices. Would different tuning (flow rates, temperatures) in this channel help in getting results closer to the other channels?

Regarding Figure 2: is this the same as Figure S1a.? I’m wondering whether for the main text it would make more sense to show the detection efficiencies for the different channels at which they have been finally operated during the ambient measurements.

Purely technical issues:

Figure 1 might benefit from some more explanatory figure caption. What do the colours in the figure indicate?

In the supplemental material preceding equations S1 and S2 it’s written “Eq. Eq. (S1)” and “Eq Eq. (S2)”. One “Eq.” can be removed there.

The article presents a new instrument consisting of five water CPCs in parallel (CPC-battery) intended for fast measurements of the particle size distribution in the 1-3 nm size range.

The study is rigorously done, including laboratory and field testing, inversion development and error analysis. The writing is very clear and easy to follow. I do not have major concerns with the study, but the authors could consider the following minor comments/questions.

Minor comments/questions:

-were there any counts from homogenous nucleation at the channel(s) with lowest cut-off size and how was that monitored?

-Figure 2. I would find it more useful that the main article would show those curves that were used for the final CPC Farm instrument, instead of different temperature settings for one channel.

-In Fig 6c there is a lot of white, where the PSD did not record any counts (I assume), meaning it underestimated by a lot. Would it make sense to assign a very dark blue color to those points, so that it is visually easier to read the plot? Currently the red points which are due to noise in PSD stand out from the white background, although that is not the main point of this plot.

-row 33. I don’t really get how the undetected NPF events could influence the growth of existing particles to larger sizes

-row 55. This is major limitation/uncertainty in some instruments.

-row 135. Can this be generalized to other environments than boreal forest where the observations in Riipinen et al. 2009 were made and is there newer studies on the hygroscopicity of 1-3 nm particles?

-Chapter 3.4. could changing particle composition also play a role in why some events/plumes are better detected than others or why the ration between the different instruments varied?