AT2020 or AT2035 on bus power? One accepts 11–52 V, the other only 48 V — and the difference is silent
A bus-powered interface generating phantom power from 5 V USB does not always reach a full 48. The AT2035's datasheet accepts anything from 11 V. The AT2020's asks for 48 and says nothing about what happens below it.
Signal Chain Compatibility Matrix
Select gear and parameters for live gain, impedance and SPL calculations.
Input Equipment Setup
PASS: Optimal Signal Chain
Requires minimum +60dB of clean pre-amp gain. Most budget entry interfaces max out at +50dB to +56dB and introduce hiss at max pot. An inline preamp booster (Cloudlifter CL-1 / FetHead) is strongly recommended.
The interface preamp provides sufficient headroom.
$538
$405
-25%Tonal & acoustic characteristics: Flat, wide-range frequency response for clean music and speech. Smooth, warm acoustic character with air-suspension shock isolation.
Technical advisory: With +69dB max gain, the Gen 4 Solo can drive demanding dynamics like the Shure SM7B natively without requiring an external inline booster.
Physical Interconnect Topology
Shure SM7B
Direct Cable Connection (No Intermediate Box)
1x Shielded Balanced XLR Cable (3-Pin XLR Male to XLR Female)
Focusrite Scarlett Solo (Gen 4)
Host: Apple MacBook Pro (M1/M2/M3/M4)
Short answer: both work on this interface at normal speaking level. But only one of them has a manufacturer's promise about what happens when the supply sags — and it is not the cheaper one.
The two microphones are close relatives and the specification sheets look nearly identical, until you reach the last line:
| | AT2020 | AT2035 | |---|---|---| | Sensitivity | −37 dBV (14.1 mV) | −33 dBV (22.4 mV) | | Gain needed here | 42 dB | 38 dB | | Max input SPL | 144 dB | 148 dB (158 with pad) | | Self-noise | 20 dB SPL | 12 dB SPL | | Phantom power | 48 V DC, 2 mA | 11–52 V DC, 3.8 mA |
At 85 dB SPL against the EVO 4's +16 dBu clip point, the AT2020 asks for 42 dB out of the 58 the interface has. Level is not the problem here. Voltage is.
Why this matters on a bus-powered box. A USB interface has 5 V to work with and generates its phantom rail with a step-up converter. Manufacturers design for the standard, but the rail is a budget item like any other, and it is shared: on a two-channel interface, two condensers draw from it at once. The published number is the target, not a measurement of your unit on your cable on your laptop port.
A microphone rated 11–52 V is being told by its maker: whatever comes out of that converter, we have tested it and the mic works. A microphone rated "48 V" has been promised nothing below 48. In practice a starved electret does not switch off — its polarising voltage falls, and with it the output and the headroom. It gets quieter and distorts earlier, gradually, with no indicator anywhere. That is the worst failure mode there is: one that looks like your voice, your room, or your technique.
The gem, and it is in Audio-Technica's own PDF. Open the AT2035 datasheet and read the "Operation and Maintenance" paragraph. It says: "The AT2020 requires 11V to 52V phantom power for operation." Wrong model name, in the AT2035's own document — a template left over from the cheaper microphone's sheet. Meanwhile the AT2020's own datasheet says 48 V in the same paragraph and again in the spec table. If you searched for this the obvious way, you could quite reasonably come away believing the AT2020 accepts 11 V. It does not, per its own sheet. Two documents from one manufacturer disagree about one microphone, and the specification table is the one to trust.
What this page cannot tell you. Our schema has no field for phantom voltage — not for what a microphone needs, not for what an interface delivers. So the green "phantom power OK" verdict above means only "this interface has a +48 V button". It is not a measurement, and on this particular pair it is not the whole question. If you are buying for a bus-powered interface, an iPad, or a USB hub, the AT2035's 11–52 V range is worth real money, and it is the line in the specification nobody reads.