High Frequency Board Manufacturing on Rogers Material
Fifth generation mobile networks, satellite links, millimetre wave radar and automotive radar have pushed high frequency circuit boards from a specialist item into a mainstream requirement. A high frequency board differs from an ordinary FR4 board in the material, in the processing and in the way quality is controlled, and the demands on the fabricator are correspondingly higher.
Where the Performance Is Decided
Two properties decide whether a high frequency board meets its specification. The first is the stability of the dielectric constant, because a variation in it shifts the phase and the impedance of every transmission line on the board. The second is the accuracy of the transmission line geometry, because the width and the spacing set the characteristic impedance, and a deviation in the width shows up as a reflection and as additional insertion loss. The two work together, and controlling only one of them does not produce a predictable board.
The Materials and the Range
Rogers materials across the range are processed, covering different dielectric constants and thicknesses for radio and microwave work, along with other common high frequency laminates where the customer requires them. Boards below the millimetre wave band are supported, including twenty four gigahertz and seventy seven gigahertz automotive radar boards, millimetre wave communication boards and radar antenna boards.
A hybrid stack-up is often the economical answer. The layers that carry the high frequency signals use the high frequency material and the remaining layers use FR4, so that the performance is where it is needed and the cost is not paid across the whole construction. Impedance controlled boards are produced with microstrip, stripline and coplanar waveguide structures, together with the test coupon at the panel edge and the report that goes with it.
What the Process Has to Control
The material is sensitive to temperature and humidity, and its storage condition affects both the dielectric constant and the dimensional stability. High frequency material is therefore kept under controlled storage, and it is baked according to the specification before it enters the line, so that the panel that is processed is in the state the design assumed.
Line width accuracy comes next. The conductor of a high frequency transmission line determines the impedance directly, so the exposure is carried out on high precision equipment, the etching parameters are controlled closely, and the region carrying the high frequency signals is inspected specifically rather than relying on a general check. Structures such as a coplanar waveguide, where the spacing to the adjacent ground is as important as the width, are controlled in the same way.
The dielectric thickness is the third item. It is the parameter that sets the impedance of a microstrip or a stripline, so the thickness after lamination is measured by sampling on each batch and compared with the stack-up, and on a hybrid construction every layer is included in that measurement rather than only the high frequency layers.
The surface finish is the fourth. Different finishes differ in conductivity, flatness and cost, and the choice follows the frequency band and the performance requirement: immersion gold suits the higher bands, while OSP suits a design where cost matters more than the last fraction of loss.
Where These Boards Are Used
Automotive applications include millimetre wave radar boards, vehicle antenna boards and advanced driver assistance radio boards in the twenty four, seventy seven and seventy nine gigahertz bands. Communication equipment includes base station power amplifier boards, filter boards and antenna element boards operating below six gigahertz and in the millimetre wave band. Security radar covers speed measurement, access control and presence detection at ten and twenty four gigahertz, and satellite communication covers receiver boards, phased array antenna boards and radio front ends across a number of bands. Test and measurement equipment uses high frequency boards in instruments, signal sources and network analysers.
Design and Production as One Exercise
The choice of material, the stack-up, the width calculation and the impedance control are design decisions that depend on what the fabricator can hold, so having both under one roof removes the argument about which side is responsible for a result. The designer and the fabrication engineer confirm the stack-up together, review the manufacturability before the artwork is released, and adjust the impedance parameters with the process figures rather than with a generic table.
A two layer high frequency prototype is normally delivered in five to seven working days, and a multilayer or hybrid board in seven to twelve, with the caveat that a material not held in stock adds its own purchasing lead time. Impedance measurements are made with a test coupon and a report is issued, and for the highest frequencies the customer is advised to confirm the performance with a network analyser at system level, because the measurement uncertainty grows at those frequencies. Our PCB manufacturing group runs the process, PCB design and layout supports the design, and quality management holds the records.
Why a Hybrid Construction Helps
`r`n
The cost of a high frequency board is dominated by the laminate, and the laminate is only needed where the high frequency signals travel. A construction that uses the high frequency material for the layers that carry those signals and ordinary FR4 for the rest keeps the electrical performance where it is required and removes the premium from the layers that do not need it. The saving is largest on a board with many layers and a small number of radio frequency paths.
`r`n
The hybrid approach has its own requirements, since the two materials have different dielectric constants and different mechanical behaviour, so the lamination parameters and the stack-up symmetry have to be planned for the combination rather than for either material alone. Where the design allows it, keeping the high frequency layers adjacent to a ground plane of the same material simplifies the calculation and makes the impedance easier to hold.
What to Send With the Order
`r`n
The Gerber data, the layer count and the material preference are the essentials, together with the frequency band the board will operate in and the impedance values that have to be held. Stating the band matters because it narrows the material choice and the surface finish without a long discussion, and stating the impedance requirement for each structure allows the width table to be calculated against the actual construction.
`r`n
Where a stack-up already exists it is supplied as well, since the dielectric thicknesses in it determine the widths. Where the requirement is still open, describing the signals that the board has to carry, the loss that can be tolerated and the space available is enough for a construction and a material to be proposed.
`r`n
For a hybrid board it also helps to say which signals are high frequency and which are not, because that decision is what allows the expensive material to be confined to the layers that need it rather than spread across the whole board.
FAQ
How much more does a high frequency board cost? The main factor is the price of the laminate, which is well above FR4, and the processing adds to it. A hybrid construction is the usual way to reduce the total.
Can the impedance be measured? Yes, with a coupon at the panel edge measured by time domain reflectometry and a report supplied with the boards.
How small an order is accepted? A few boards can be produced for a prototype, though the engineering and material cost is spread over fewer units.
