RF coupler specifications for a German design review are thinner than the review expects. Of the 17 couplers Walsin Technology (Walsin) publishes, none states directivity or isolation, and the coupling figure covers 43 to 57 percent of the frequency span each part claims. We read the header row of all 203 export files.
Venatech Limited, an authorised Walsin distributor in Hong Kong, holds all 203 parametric export files from Walsin, 15 of them in the RF folder. This post is written from the coupler file, the antenna switch file and the chip attenuator file, read as raw cells. Every part number below is real. We hold no Walsin RF datasheet of any kind, so every figure is the export’s own value and a shortlist aid, not a specification. A German power-detection or output-monitoring design asks for coupling, directivity and isolation. Here is what the data holds for each.
Which RF coupler specifications does Walsin actually publish for a German design review?
One of the three, and only on couplers. We read the header row of all 203 export files and searched for directivity, isolation, coupling, return loss and standing wave. Directivity appears in 0 files. Coupling appears in 2 files, which are byte-identical twins of one 17-row coupler list. Isolation appears in 3 files, none of them a coupler file.
| Parameter a coupler decision needs | Column name in the export | Files carrying that header (of 203) | Parts that populate it (of 457 RF parts) | Match / mismatch for a coupler requirement |
|---|---|---|---|---|
| Coupling factor | Coupling in BW |
2 (one list, stored twice) | 17 of 17 couplers | Match on presence; mismatch on format, see below |
| Directivity | none | 0 | 0 | Mismatch — not published on any RF part |
| Isolation | Isolation(dB) |
3 | 9 antenna switches, 0 of 17 couplers | Mismatch — a switch field, absent on couplers |
| Return loss | none | 0 | 0 | Mismatch — nothing to filter |
| Insertion loss | Insertion Loss (dB) |
6 | 17 of 17 couplers | Match on presence; column naming is covered in our insertion loss post |
The RF folder stores every export twice, once as FRESH_rf_* and once as Walsin_*. The naive row total is 914. The real total is 457, and every count in this post is of distinct part numbers.
How is the coupling cell written, and how many rows give a usable value?
Seventeen of 17 hold something, and 13 of 17 give a value you can use in a calculation. We parsed all 17 Coupling in BW cells and masked every number to see the cell’s template: 7 distinct templates across 49 lines of text. The column is called a bandwidth figure, yet four different cell shapes live in it.
| Cell form | Parts | Real example | Unit stated | Frequency stated | Match / mismatch against “dB at a named frequency” |
|---|---|---|---|---|---|
| Range with dB and an MHz window, 3 or 4 lines | 12 | RFCPL1608070P2T |
yes | yes | Match |
| Seven point values, no unit | 1 | RFCPD1608050PS5T92 |
no | yes, 7 points | Mismatch — unit missing, and no continuous window |
One value with tolerance, 10.0 ±1.0 dB |
1 | RFCPL1810B2450T |
yes | no | Mismatch — frequency implied by the part’s 2400 to 2500 MHz band |
| Bare number | 3 | TFCPL0605B2450T, TFCPL0605B24508Q1C, RFCPL1806B2450T |
no | no | Mismatch — no unit, no window, no tolerance |
13 of 17 rows carry the unit dB in the coupling cell and 4 do not. The 4 without it include both Thin Film Coupler parts. Spacing varies as well: dB@ and dB @ are both used, and the tilde sits with or without spaces. A string match on the column reads these as different values. Our earlier post on Walsin’s coupler series covers the series split and the 2.4 GHz coupling trade-off. This post is about the column.
How much of the stated band does the coupling figure cover?
Between 42.8 and 56.9 percent on the 12 banded parts. We summed the width of every coupling sub-band on each part and divided by the span in the Frequency column: 852 to 1,134 MHz of a 1,992 MHz span on the 698 to 2690 MHz parts. The rest of the stated band carries no coupling figure at all. We have not seen this counted anywhere.
| Coupler group | Parts | Frequency column span | Coupling stated across | Share of span | Last MHz with a coupling figure | Match / mismatch against a “whole band” requirement |
|---|---|---|---|---|---|---|
| 698 to 2690 MHz, mid window from 1427.9 MHz | 5 | 1,992 MHz | 1,134.1 MHz | 56.9% | 2620 | Mismatch — 70 MHz tail unstated |
| 698 to 2690 MHz, mid window from 1710 MHz | 5 | 1,992 MHz | 852.0 MHz | 42.8% | 2620 | Mismatch — 282 MHz less than the group above |
700 to 2700 MHz, RFCPL1608070PM9T16 |
1 | 2,000 MHz | 969.0 MHz | 48.5% | 2700 | Mismatch — gaps between sub-bands, no tail gap |
617 to 2690 MHz, RFCPL1608070PFT |
1 | 2,073 MHz | 933.0 MHz | 45.0% | 2620 | Mismatch — 70 MHz tail unstated |
699 to 2690 MHz, RFCPD1608050PS5T92 |
1 | 1,991 MHz | 7 point values | not computable | 2690 | Mismatch — points, not windows |
Coupling stops at 2620 MHz on 11 of the 12 banded parts while their insertion loss is stated to 2690 MHz. A design that samples power near the top of the band is outside the stated coupling window on those parts.
Are the 10 couplers on the largest Walsin shelf the same part?
No, and the file shows how. We compared the 10 parts that share the 698 to 2690 MHz band, the 1608 size, 6 pins and 50 ohm impedance, field by field, across both the coupling and insertion loss cells: they split into 4 variants on two fields that look like nothing. Our earlier shelf count treated the 10 as ten second-source candidates. We have not seen this split published anywhere.
| Variant | Parts | Mid-band coupling window | Insertion loss wording | Match / mismatch within the 10 |
|---|---|---|---|---|
| A | RFCPL1608070P08Q1C, RFCPL1608070P0T, RFCPL1807P2S0T62 |
1427.9 to 2025 MHz | max. stated |
Match to each other |
| B | RFCPL1608070P2T, RFCPL1608070P28Q1C |
1427.9 to 2025 MHz | no max. |
Mismatch with A on whether 0.20 to 0.25 dB is a limit |
| C | RFCPL1608070P1T, RFCPL1608070P3T |
1710 to 2025 MHz | max. stated |
Mismatch with A on the window start, 282.1 MHz later |
| D | RFCPL1608070P38Q1C, RFCPL1608070PS0T62, RFCPL1608070P18Q1C |
1710 to 2025 MHz | no max. |
Mismatch with A on window and wording |
All 10 agree on the low window (23.0 to 27.0 dB at 698 to 915 MHz), the high window (22.5 to 27.5 dB at 2300 to 2620 MHz) and the insertion loss numbers (0.20, 0.22 and 0.25 dB). They differ on where the middle coupling window starts and whether the insertion loss is stated as a maximum. 5 of the 13 parts with a banded insertion loss give no maximum, so the cell does not say whether 0.20 dB is a limit or a typical value. Add a column for your incumbent part, mark each row, and the split decides which of the 10 you can offer.
Where do isolation and attenuation live in the RF line instead?
Isolation lives on the 9 antenna switches, and attenuation on 23 chip attenuators in a different folder. We opened Walsin_Antenna_Switch.xlsx and the chip attenuator file and read every cell: all 9 switch rows populate isolation, as 3 to 6 lines of text each, and all 23 attenuator rows state attenuation in a column named Resistance.
| File | Rows | Parameter | Where it sits | Match / mismatch against its own column name |
|---|---|---|---|---|
Walsin_Antenna_Switch.xlsx |
9 | Isolation, dB minimum by sub-band | Isolation(dB), 9 of 9 rows |
Match on column; 2 rows add a port-routing label inside the cell, for example RFASWA630PTF0E has none and RFASWR681CTF0G has one |
Chip_Attenuator_Convex_type__1of1.xlsx |
23 | Attenuation, 0 to 20 dB | Resistance |
Mismatch — decibels in a resistance column |
| same | 23 | Port impedance | Tolerance |
Mismatch — 50 ohm on 22 rows, the word Jumper on the 0 dB part |
| same | 23 | Attenuation accuracy | TCR |
Mismatch — eight values from plus or minus 0.1 to plus or minus 2.5 dB |
22 of 23 attenuators are usable steps. The 0 dB part WA04P000XPTL is a jumper. Neither the attenuator file nor any other RF file states frequency range or return loss for those parts, which our chip attenuator cross-reference post explains in full. The antenna family is separate: its Voltage Standing Wave Ratio column is filled on 36 of 44 rows.
What should a German design review ask for before a coupler goes on the BOM?
Four things, none of which the export supplies. We counted 0 parts of 457 carrying directivity and 0 of 17 couplers carrying isolation. A directional coupler without directivity cannot be judged on how much reflected power leaks into your forward-power reading.
- Directivity at your operating frequency. Ask for the Walsin datasheet for the exact part number. We hold none for RF.
- A maximum or typical statement on insertion loss. 5 of the 13 banded parts do not say which.
- Coupling over your whole band, including the top 70 MHz. 11 of 12 banded parts stop at 2620 MHz.
- The mid-band window start. 1427.9 or 1710 MHz decides whether a part matches your incumbent.
Our free cross-reference tool states differences from your original part instead of hiding them. The wider RF line is on the RF products page, and what we hold in Hong Kong is on ready stock as a dated snapshot of 1,125 part numbers as at 23 September 2026, not a live feed. We hold 28 Walsin datasheets, 451 pages, none for RF. Confirm every figure against both datasheets and qualify in your own circuit.
The short version
Walsin’s RF export publishes coupling on 17 couplers in 7 cell templates, and nothing for directivity in any of 203 files. Isolation sits on 9 antenna switches only. Coupling covers 42.8 to 56.9 percent of the stated span on 12 banded parts, and stops at 2620 MHz on 11 of them. The 10 parts on the biggest shelf split into 4 variants on the mid-band window start and on whether insertion loss is a maximum. Attenuation is stored in a resistance column. We will not tell you a coupler matches until your directivity requirement meets a datasheet.
Send us the coupler line off your BOM and the band you sample, and we will return Walsin candidates field by field. And a question back: when your review asked for directivity, which document did you expect to find it in?
Frequently asked questions
Does Walsin publish directivity for its RF couplers?
No. We read the header row of all 203 export files and the word directivity heads none of them, so 0 of 457 RF parts carry it. Walsin publishes a coupling figure on all 17 couplers and an insertion loss figure, but a directivity value would have to come from a datasheet, which we do not hold for RF.
Which Walsin RF parts publish isolation?
Only the 9 antenna switches, in a column named Isolation(dB), as 3 to 6 lines of minimum decibel values by sub-band. None of the 17 couplers carries isolation. The antenna switch module file has the same column headings and returns zero rows. Isolation for a coupler port is not in the export.
How many Walsin couplers give coupling in decibels at a stated frequency?
13 of 17 carry the unit dB, and 12 of those give a range with a megahertz window. One more gives seven point values with no unit. Four rows carry a bare number or a value without the unit, including both Thin Film Coupler parts. The column holds 7 distinct cell templates across 49 lines.
Does the coupling figure cover the whole stated band?
No. On the 12 banded parts, coupling is stated across 42.8 to 56.9 percent of the span in the Frequency column. On 11 of the 12 the last coupling window ends at 2620 MHz, while insertion loss is stated to 2690 MHz. A design near the top of the band sits outside the stated coupling window.
Are Walsin’s 698 to 2690 MHz couplers interchangeable on their data?
Not on the data alone. Ten parts share the band, size, pin count and impedance, and they split into 4 variants. Five start the middle coupling window at 1427.9 MHz and five at 1710 MHz, and 5 of the 13 banded parts give no maximum on insertion loss. Compare field by field against your incumbent.
Where is the attenuation value for a Walsin chip attenuator?
In the column named Resistance, as text from 0dB to 20dB on 23 rows. Port impedance sits in the Tolerance column as 50 ohm on 22 rows, and attenuation accuracy sits in the TCR column as eight values from plus or minus 0.1 to plus or minus 2.5 dB. The 0 dB part is a jumper.
What documentation does Venatech hold for Walsin RF parts?
None at datasheet level. We hold 28 Walsin datasheets, 451 pages, covering 26 MLCC and 2 chip resistor series, and none for RF, antenna, inductor, varistor, disc capacitor or thin film. Every RF figure here is the export’s own value and is flagged for confirmation against Walsin’s own datasheet.