Published Nov 12, 2026 · Last updated Nov 12, 2026 · 5 min · IndieRF

Cascaded P1dB, LNA Into Mixer: −8.8 dBm

An HMC8413 LNA into an HMC213B passive mixer compresses at −8.8 dBm at the input, set by the mixer's 11 dBm input P1dB.

Put the HMC8413 ahead of the HMC213B and the pair compresses at −8.8 dBm (0.13 mW) at the chain input, set by the mixer. The LNA alone would take +3 dBm; the mixer's 11 dBm input P1dB behind 19.5 dB of gain is −8.5 dBm at the input, and the LNA's own limit costs only 0.3 dB more.

The Inspector's First to Compress readout names the mixer. At the −30 dBm source in the link, P1dB Headroom reads 21.2 dB.

Two data sheets, one convention

Analog Devices prints the HMC213B's compression at its input. The HMC213BMS8E data sheet, Rev. 0, page 3, Table 1, downconverter at 25 °C, 100 MHz IF, upper sideband: from 1.5 GHz to 4.5 GHz with 13 dBm of LO, conversion loss 10 dB, input IP3 21 dBm, input P1dB 11 dBm, all typical. The HMC8413 side comes from its Rev. 0 data sheet, page 3, Table 1 (0.01 GHz to 7 GHz): gain 19.5 dB and OP1dB 21.5 dBm typical, the same values the IP3 note uses. The chain runs at 2.4 GHz, inside both parts' ranges.

The LNA's P1dB is an output number and the mixer's is an input number, so move both to the chain input before comparing them. With the usual 1 dB convention, IP1dB=OP1dB−G+1\mathrm{IP1dB} = \mathrm{OP1dB} - G + 1:

  • LNA: 21.5−19.5+1=321.5 - 19.5 + 1 = 3 dBm.
  • Mixer: 11 dBm at its own RF port, which is 11−19.5=−8.511 - 19.5 = -8.5 dBm at the chain input.

Then add the reciprocals in milliwatts, each stage's limit divided by the gain ahead of it:

1IP1dB=1IP1dB1+G1IP1dB2=12.00 mW+10.141 mW\frac{1}{\mathrm{IP1dB}} = \frac{1}{\mathrm{IP1dB}_1} + \frac{G_1}{\mathrm{IP1dB}_2} = \frac{1}{2.00\ \mathrm{mW}} + \frac{1}{0.141\ \mathrm{mW}}

That is 0.50+7.08=7.580.50 + 7.08 = 7.58 per mW, or 0.132 mW, −8.80 dBm. The mixer's term is 14 times the LNA's. The output side reads the same chain from the IF port: 9.5 dB of net gain puts the cascaded OP1dB at −0.3 dBm.

Cascade sums 1/(OP1dBi Gafter,i)1/(\mathrm{OP1dB}_i\,G_{\mathrm{after},i}) at the output and converts with the same 1 dB, which is the input formula above multiplied through by the chain gain. The test for this note works the input form by hand and the engine agrees to 1e-9 dB. Both assume soft compression in every stage, so treat the result as the point where the chain starts to bend, good to a dB or so when one stage dominates as it does here.

Cascade's mixer field is Output P1dB, so the link enters the HMC213B as 11−10−1=011 - 10 - 1 = 0 dBm. Type the printed 11 into that field and the chain reads −0.3 dBm at the input, 8.5 dB optimistic.

Less LNA gain, more chain P1dB

Table 1 of the HMC8413 data sheet prints two gains for this band, 19.5 dB typical and 17.5 dB minimum, and a 19 dBm minimum OP1dB. Table 2 (7 GHz to 9 GHz) is above the mixer's 4.5 GHz RF limit, so the table has three rows: typical, minimum gain, and minimum gain with minimum OP1dB. The HMC213B prints its input P1dB at two LO drives, and each sits on its own frequency row. Every chain value below is the Cascade engine on the linked design, checked against the hand formula.

LNA gain (dB)LNA OP1dB (dBm)LNA IP1dB (dBm)Mixer at input, LO 13 dBm (dBm)Chain IP1dB, LO 13 dBm (dBm)Chain IP1dB, LO 10 dBm (dBm)
19.521.53−8.5−8.80−9.27
17.521.55−6.5−6.80−7.27
17.5192.5−6.5−7.01−7.46

Every dB of LNA gain is a dB of chain P1dB while the mixer dominates. A part at minimum gain compresses 2 dB later and pays for it in noise figure. When the LNA also lands at its 19 dBm minimum OP1dB, its term grows and costs another 0.2 dB, still an eighth of the mixer's term.

The LO 10 dBm column uses the data sheet's 1.7 GHz to 3.6 GHz row (conversion loss 10.5 dB, input P1dB 10.5 dBm), so the chain sits 0.5 dB lower. That row changes LO drive and band together, and the sheet prints nothing to separate them.

Where the budget goes wrong

When a chain compresses earlier than the budget says, the stage I find overlooked most often is the mixer.

The mixer's input P1dB is the number to look up, and it is printed. The HMC213B's 21 dBm input IP3 sits 10 dB above its 11 dBm input P1dB, near the 10 dB gap the class-A amplifier rule predicts. That rule breaks for mixers, where the gap runs from 5 to 15 dB, so read the printed P1dB before you estimate one from IP3.

The typical column at 25 °C is all the HMC213B prints for P1dB, with no minimum. A blocker at the input sees the same limit as the wanted signal, since compression follows total RF power.

References

  • Analog Devices, HMC213BMS8E, 1.5 GHz to 4.5 GHz, GaAs, MMIC, Double Balanced Mixer, data sheet Rev. 0, 2/2018, page 3, Table 1.
  • Analog Devices, HMC8413 Low Noise Amplifier, 0.01 GHz to 9 GHz, data sheet Rev. 0, 10/2021, page 3, Table 1.
  • B. Razavi, RF Microelectronics, 2nd ed., section 2.2: gain compression and cascaded nonlinear stages.

Questions? Contact

Discussion

No comments yet. Start the thread with a measurement, a correction, or a worked example.