Introduction
Most commercial surface mount RF switches on the market do not specify how well they work at HF frequencies. In reality, most are likely to work quite well, despite most switches having a minimum input frequency of 10 or 20 MHz. This post characterizes a number of switches available on the market in the SC70-6 / SOT363 footprint between 500 kHz and 100 MHz. I am primarily interested in the insertion loss and isolation of the switches.
Measurement setup
Individual switches were soldered onto SC70-6 to DIP breakboards available on Amazon. Photographs showing a close-up view of a soldered part and the full breakout board are shown below.
The devices tested here all have the same pinout, shown below. Some devices allow a single control logic signal in addition to complementary logic signals; all devices were tested with complementary control logic signals, i.e., if one control line was driven high, the other was driven low.
I used a measurement jig to enable devices under test to be switched in and out. The control lines were driven at 3.3V by a microcontroller. A picture of the test jig is shown below. A 100 nF capacitor was inserted between the RF pins and the SMA connectors.
A calibrated NanoVNA was used to measure the insertion loss between the common RF connection (RFC) and the two output RF connections RF1 and RF2; the RF output not being tested was terminated with a 50 Ohm load.
Switches under test
| Part # | LCSC | Digikey | Mouser |
|---|---|---|---|
| AS179-92LF | $1.36 | $1.06 | $1.06 |
| XA17-G4K | $0.54 | NA | NA |
| TPHWS314 | $0.14 | NA | NA |
| TPE4259-63 | $0.13 | NA | NA |
| HX4259-63-ST | $0.15 | NA | NA |
| PE4259-63 | $0.62 | $0.79 | NA |
| PE42421-SCAA-Z | $1.51 | $0.79 | $0.79 |
| MASWSS0115TR-3000 | $1.97 | $1.32 | $1.32 |
| MASWSS0179TR-3000** | $0.48 | $1.06 | $1.05 |
(NA = Not Available. **Different footprint, added for reference. Prices are for minimum order quantity, checked on 2 April 2026)
Insertion Loss and Isolation
All the devices continued to perform well below their minimum rated frequency. The amateur radio bands are highlighted greenish in the plots below.
Note that the absolute value of the insertion loss includes the loss of the measurement jig and breakout board and is not representative of the SMD device’s insertion loss. However, since all devices shared the same measurement jig and breakout board, they can be compared relative to each other.
Linearity
In order to obtain measurements of linearity at powers up to 30dBm I inserted a QB-101 amplifier between the signal generator and the switch. I added 40dB of attenuation between the switch and power meter to avoid overloading the power meter. The QB-101 amplifier started to compress at the higher end of the range, as shown in the curve below, but its P1 point is higher than 30 dBm so did not affect the conclusions of this test. All linearity measurements were made with a 2 MHz signal source.
The linearity of the switches is shown in the plot below. The switches fall into four groups.
- The switches with the worst linearity, corresponding to a P1 of roughly 8 dBm, were the MASWSS0115TR-3000 and AS179-92LF.
- The next best switches, TPHWS314 and TPE4259-63, have a P1 of roughly 18 dBm.
- The next best switches, XA17-G4K and HX4259-63, have a slightly better P1 of roughly 20 dBm.
- The best switches, PE4259-63 and PE42421SCAA-Z, have a P1 of > 30 dBm,
Conclusions
A table comparing the specified and measured performance of the switches tested is provided below. I also include whether the switch is capable of being controlled by a single logic line (i.e., it does not need complementary logic control lines) and the switch behavior (all switches behave the same except PE42421-SCAA-Z, which is reversed).
| Part # | LCSC price | Specified P1 [dBm] | P1@2MHz [dBm] | Isolation [dB]* | Single control | Unconnected switch | RFC to pin 3 when pin 4 is HI |
|---|---|---|---|---|---|---|---|
| AS179-92LF | $1.36 | 30 | 8 | -85 | No | Float | Isolation |
| XA17-G4K | $0.54 | 30 | 20 | -79 | No | Float | Isolation |
| TPHWS314 | $0.14 | 30 | 18 | -78 | No | Float | Isolation |
| TPE4259-63 | $0.13 | 31 | 18 | -76 | No | Float | Isolation |
| HX4259-63-ST | $0.15 | 33.5 | 20 | -75 | Yes | Ground | Isolation |
| PE4259-63 | $0.62 | 33.5 | >30 | -64 | Yes | Ground | Isolation |
| PE42421-SCAA-Z | $1.51 | 33.5 | >30 | -65 | Yes | Ground | Insertion loss |
| MASWSS0115TR-3000 | $1.97 | 25 | 8 | -48 | No | Float | Isolation |
| MASWSS0179TR-3000** | $0.48 | 16 | -48 | No | Float | Insertion loss |
(*At 1MHz. **Different footprint, added for reference.)
Almost every switch tested works significantly better than the MASWSS0179 version used in the T41 V12 radio. The optimum parts from a cost and isolation performance perspective are provided in the table below. They are Asian brands not available through Digikey or Mouser.
| Part # | LCSC price | Specified P1 [dBm] | P1@2MHz [dBm] | Isolation [dB]* | Single control | Unconnected switch | RFC to pin 3 when pin 4 is HI |
|---|---|---|---|---|---|---|---|
| TPHWS314 | $0.14 | 30 | 18 | -78 | No | Float | Isolation |
| TPE4259-63 | $0.13 | 31 | 18 | -76 | No | Float | Isolation |
| HX4259-63-ST | $0.15 | 33.5 | 20 | -75 | Yes | Ground | Isolation |
The AS179-92LF part has the best isolation performance and is available from US suppliers, but is also significantly more expensive and has the worst linearity at HF frequencies. If linearity is important, then the PE4259-63 and PE42421-SCAA-Z offer the best linearity with good isolation and are avaliable at Digikey.