Contact Us
  • Home
  • BLOG
  • Why Does Conformal Coating Detune an RF Board?

Why Does Conformal Coating Detune an RF Board?

A conformal coating detunes an RF board because it puts a dielectric where air used to be. On a microstrip line or a printed resonator, a large share of the field sits in the air above the copper. A film with a dielectric constant of 2.6 to 4.4 laid over that copper raises the effective permittivity of the line, slows the wave down, and pulls every resonant structure on the board to a lower frequency. A filter centred at 8 GHz does not stay at 8 GHz, and the shift is permanent because the film stays on the board.

The size of the move is set by three things and none of them is the brand on the drum: the dielectric constant, the thickness, and how much of the field the film covers. Measured on a coated filter assembly after 250 hours of damp heat, the centre frequency moved 0.07 GHz at 4 GHz, 0.11 GHz at 8 GHz and 0.18 GHz at 12 GHz. That is between 1.4 and 1.8 percent of centre frequency, and a narrow band filter does not have 1.8 percent of margin to give away.

Why a thin film moves a resonator so far

Frequency on a transmission line goes as one over the square root of effective permittivity. Air is 1.0 and every coating on the shelf is between 2.6 and 4.4, so the film does not need to be thick to matter, it needs to be where the field is strong. On a microstrip that is the 50 to 150 um directly above the copper, which is exactly the thickness a spray or a dip produces without trying.

The same board in stripline behaves differently. The field is enclosed in laminate and the outer film only sees the fringes, so the shift is smaller and often lost in the tolerance. This is why two products with the same coating call-out can react in completely different ways, and why a number taken from another project is not usable as a prediction.

What the numbers look like

Dielectric constant at 1 MHz for the common chemistries: parylene N 2.65, parylene C 2.95, acrylic 2.7 to 3.2, epoxy 3.1 to 4.2, silicone 3.1 to 4.0, polyurethane 3.8 to 4.4. Dry thickness: parylene 12.5 to 51 um, acrylic, epoxy and urethane 25 to 127 um, silicone 51 to 203 um. Silicone has a low dissipation factor and the thickest film of the group, so it loads the line more than a thin parylene layer even though it is the better moisture barrier. Polyurethane has the worst loss of the list, at 0.068 to 0.074, and it is the one that shows up first when a coated assembly loses gain.

Loss is the second half of the problem. On the same filter, insertion loss at 4 GHz was 0.60 dB bare, 0.47 dB with a 1 um low loss coating, and 2.03 dB with parylene C at 12.5 to 51 um. At 8 GHz the three were 1.58, 1.04 and 4.62 dB. Parylene C has a dissipation factor of 0.013 against 0.003 to 0.006 for silicone, and at 8 GHz that difference is measured in decibels, not in percent.

The part that happens after the coating is dry

Every polymer takes up water. Water has a dielectric constant near 80, so a film that absorbs moisture raises its own permittivity and its loss over weeks in the field. The 250 hour figures above are not as-coated figures, they are what the assembly measured after damp heat, and that is the state the radio lives in for years.

Thickness spread adds the rest. A spray coat on a board with tall components varies across the panel, and each 10 um of extra film over a resonator moves it a little further. Two boards from the same batch can therefore sit at different frequencies, which is the failure that is hardest to argue about because both meet the drawing.

Where builds go wrong

Coating the matching network to protect it. The network sits next to the connector and gets the heaviest coat, which is precisely where the field is most exposed. The area that needs protection and the area that must stay bare are not the same area, and the difference has to be on the drawing as a boundary, not as a verbal instruction at the spray booth.

Qualifying on a coupon. A bare coupon coated in the same run tells you the thickness. It does not tell you the shift, because the shift depends on the copper pattern around the film. The only useful check is S21 and S11 on the real assembly before and after coating, at temperature, on more than one board.

Treating IPC-CC-830 as the whole specification. That standard qualifies a coating material through dielectric constant, dissipation factor, moisture and insulation resistance and the rest of its test list. It does not fix your thickness, your coverage boundary or your cure. A purchase order that says coat per IPC-CC-830 leaves the three variables that decide the frequency open.

How we hold the coverage

At SUPERB Assembly the coating call-out is a material, a dry thickness range and a keep-out map, and the map is checked against the actual board before the first panel is sprayed. Connectors, test points and the areas named by the customer are masked with tape and plugs, the coat goes on by spray, dip or selective equipment depending on the mix on the board, and cure follows the material data sheet rather than a line default. Dry film thickness is measured on coupons that travel with the panel plus fixed points on the board, because a coupon alone hides the thin areas.

RF validation stays with the customer's lab or a partner RF lab. Our line tests DC, function and the customer fixture, and the coating record travels with the serial so a shift found in the lab can be traced to a thickness and a material lot. The boards themselves come from RF PCB work with controlled impedance and the same laminate family across builds, assembled as shown on the RF filter board page.

What to send with the question

The stack-up with copper weights, the frequency band and the bandwidth of the structures that matter, the coating material and the thickness range, the areas that must stay bare, the environment the board sees in the field, and the test you will use to accept the coated assembly. With those six items the answer is a coating specification with a boundary and a thickness, and the frequency question is settled before the first production panel rather than after.

Need a quote or a review of your build against these limits? pcba@superb-tech.com · WhatsApp +86 133 9608 1443