FAQs: Flex-rigid

Andreas Schilpp, our expert regarding the flex-rigid technology, answers your technical questions from the webinar "Flex-rigid Design Guide Part I and Part II".

Your question is not answered here?
Please do not hesitate to ask Mr. Schilpp via email!


What are the differences of LCP compared to Polyimide?

Liquid Crystal Polymer (LCP) foils have lower moisture intake and better dimensional stability compared to Polyimide. The availability of LCP foil is worse and the price is higher compared to Polyimide. Avalilable only for prototype samples.

Can you comment on the stack-up use of RA copper/ED copper in static/dynamic installation?

For static applications (IPC-2223 Use A – „flex-to-install“) or dynamical bending applications with little number of bending cycles or uncritical bending conditions ED copper (Electro-Deposited) is standard. RA copper (Rolled-Annealed) is used in dynamic applications (IPC-2223 Use B – „dynamical bending“), applications with several billions of bending cycles are known.

Please note:

  • RA copper has a preferred direction “MD” = “machine direction”. In the other direction the bending capability is remarkably worse
  • In case of flex-rigid with flex layer outside generally also the flexible area is plated – so this copper quality is ED-copper
  • More copper qualities are specified in IPC-4562 standard

Spacings and Creepage - Can we treat the flex section (Polyimide) as standard FR4?

No! FR4 materials have PTI / CTI level 3 up to 0 depending on the product (i.e. Nanya NP140 – CTI3 / NPG – CTI 2 / Panasonic R-1655W – CTI 1, just to list some options at least)

But Polyimide, i.e. DuPont Pyralux AP or Panasonic R-F77x only has CTI level 4.

On the other hand almost all designs do not have free copper on the flex layer without a barrier in between like soldermask or coverlay. In addition, in case of flex inside the stackup all copper is completely covered by rigid material in the rigid sections or coverlay in the flexible sections.

Regarding isolation, is Polyimide in the same material group as FR4? (IIIa/b)?

Referring to the following table: FR4 dependent on type is then II or IIIa, but Polyimide is IIIb.


Gigabit signal – is it advisable to use flex-rigid?

Yes, many USB3 cameras use Flex-Rigid circuit boards (USB3.1 Gen.2 supports a data rate of 10 Gbps). Important for this is a carefully tuned design with specified impedances. The USB3 specification requires differential lines with 90Ohm characteristic impedance. We can calculate the necessary design parameters for you, suitable for an adapted layer stack. Usually polyimide cores with a thickness of 75µm or 100µm are used.

Rigid-flex technology offers the following advantages for high data rates:

  • the flex material polyimide is a low-loss material
  • the copper lamination on polyimide has a flat backside which is favourable for high frequency (very flat treatment)
  • Rigid-flex allows the elimination of plugs, which are basically to be regarded as weak points.

Is there a lower temperature limit for rigid-flex PCBs?

For low temperatures, there are no specifications or limitations on the base materials used for rigid-flex PCBs. Our flex-rigid technology has been tested by UL according "Standard UL796 F Item 5.10 Cold Bend Test" and has passed these tests at -20°C.


Bending of FR4 Semiflex – is it allowed in both directions?

As described in the design rules FR4 Semiflex bending is only allowed with milled area inside. To achieve an S-shaped bending two bending areas, on on TOP and one on BOTTOM are necessary with a rigid part in between.

What is the minimum annular ring for vias in flex-rigid technology?

In general there is no difference to rigid circuit boards. The values are defined in IPC-6013, Table 3-11 and dependant from IPC-class specification.

For example: according IPC class 2 up to 90° breakout of hole from land is allowed, for IPC class 3 the annular ring has to be minimum 50µm for external layers and 25µm for internal layers. The use of teardrops basically improves the manufacturability, see Figure 3-6.

Is it possible to get stack-up drawings from you?

Yes, we like to share our standard stack-up drawings with you or even individual stack-up drawings meeting your special requirements, for expample for impedance controlled designs.

Some standard stack-ups you can find here.

Which software do we need to design flex-rigid?

Flex-rigid layouts are possible with many EDA programs, from Altium to Zuken. The performance of the individual systems is different, modern programs offer various support for board planning, sectional stackups up to 3D bending and installation simulations and collision testing.

The dimensional definition of the flexible and rigid areas is done e.g. on an additional mechanical help layer (Infolayer).

With more powerful systems that have a "Multi Stackups mode", this is more convenient in the "Board Planning Mode":

This definition is necessary in order to be able to create the programs for the mechanical separation of rigid and flexible areas.

What is the maximum copper thickness with flex-rigid?

Flexible base material Polyimide is available with different copper layer thicknesses. The standard thickness is 18µm and goes up to nominal 70 µm. Due to the machining processes these thicknesses change, the tolerances according to IPC-6013 for inner layers apply.

Note: For plated outer layers IPC-6013C Table 3-19 applies.

Flex inside or outside – what are the advantages/disadvantages and your recommendation?

Put in very simple terms, we recommend:

  • 1 Flex layer ⇒ outside as 1F-xRi ⇒ favourable
  • 2 or more flexible layers ⇒ inside than xRi-yF-xRi ⇒ more expensive
  • dynamic application with many bending cycles ⇒ inside + RA copper

In individual cases, e.g. with sensitive and very fast signals over the flex range, an asymmetrical design can offer 2F-xRi signal technical advantages. However, the production of such a setup is more complicated than a symmetrical setup xRi-yF-xRi. With regard to the permitted minimum bending radius, the 2F-xRi setup offers no advantage over an xRi-2F-xRi, but at the same time there is a risk of bowing and twisting due to the asymmetry in the stack-up.

Is there a possibility of a bending radius of 3 mm with Semiflex and what has to be regarded for this?

According to design rules the standard bending radius is 5mm respectively 4 mm with adapted stack-up. Smaller bending radii are possible in general but have to be designed and qualified for every application individually. We would like to assist you.

Please note that the specifications in the design rules "FR4 Semiflex" regarding bending direction and bending tool must be regarded.

How much is the maximum ampacity in the flexible or semiflexible area with same load on all conductors?

This question cannot be answered in this way. For the current carrying capacity the IPC-2152 "Design guideline for the determination of the current carrying capacity of printed circuit boards" is to be applied. The balance between heating by current (current strength, copper cross-sections) and heat dissipation (thermal conductivity and thickness of materials, via constructions, copper layout). Other important parameters are the maximum permissible excess temperature and the layer structure!

Flex-rigid stack-up: Should a flexible layer be considered as an external layer (subject to pollution level) or as an internal layer (pollution level 1)?

That depends on the stack-up whether Polyimide is used asymmetrically outside (1F-xRi / 2F-xRi) or symmetrically inside (xRi-yF-xRi), mostly based on mechanical and / or electrical constraints.

But at the same time almost all designs do not have free copper on the flex without a barrier in between like soldermask or coverlay. Moreover, in case of flex inside the stackup all copper is completely covered by rigid material in the rigid sections or coverlay in the flexible sections.

From a current carrying perspective on the flex layer: Max Copper weight we can consider is 2oz after plating?

Flexible layer outside is plated (starting 1oz base copper), flexible cores inside are available with 2oz and they are NOT plated. Tolerances according IPC-6013C Table 3-18 and Table 3-19 have to be considered.

Can we use IPC-2152 for the current carrying estimates?

Yes, I don´t know any other specification for that. For longer flex tails, however, it must be taken into account that the current carrying capacity is noticeably reduced by less material mass and thus less heat dissipation in the material. See also IPC-2125 A4.1

How often can the Semiflex connection be bent back and forth?

According to our design rules FR4 Semiflex typically 10 times bending back and forth is no problem.

Can you view and process Altium CAD data directly?

No, we generally work on the basis of post-processor data, e.g. Gerber or ODD++.

Which maximum copper thicknesses are possible with Semiflex in the bending area?

According to the FR4 Semiflex design rules, the inner layers are made of base copper 18µm and the outer layers of base copper 12µm in the standard version. According to IPC-6013, the minimum values for the copper layer thickness on the inner layer are 11.4µm and on the outer layer 29.3µm:

For higher requirements, 35mm base copper (1oz.) can be used on the inner layer and up to 35µm base copper on the outer layer.

May non used pads / non functional pads be removed with flex-rigid?

In order to ensure reliability, unconnected viapads on the flexible layers must not be removed. Since the EDA tools cannot map this differently for different layers, we recommend to generally avoid it.


Could heat be applied to assist the bending of Semiflex?

Generally yes, but we do not recommend this. To get a noticeable softening of the epoxy resin a high temperature clearly above Tg has to be applied. But this could lead to aging and embrittlement. On the other hand we recommend optimizing the stack-up if necessary and in particular the use of a bending aid!

What about the thermal resistance of Semiflex during overmolding with plastic?

The thermal resistance of FR4 Semiflex is exactly the same as the used base material FR4 respectively FR4 with a higher Tg or T260 value. During the design of the molding tool it has to be regarded that the filling process will not load the semiflexible area which could damage this.

Can you twist Semiflex?

No, twisting produces small radii and forces that a Semiflex bending area cannot withstand.

Are there any implications in PCB soldering, or is the process the same as rigid PCB?

In general flexible and flex-rigid PCBs have to be dried prior to soldering.

What if you want to solder for example relays on the other side of the flex PCB?

You should always place and solder components on rigid parts, not on flexible parts. So use Rigid-flex technology best.

SMD assembly - are there significant tolerances for the positioning of SMD components with rigid-flex?

Würth Elektronik has defined high-quality standards (see specification in the design rules  link) and uses high-quality, adhesive-free polyimide base materials; we do not use acrylic adhesives for bonding to the rigid materials. The rigid-flex printed circuit boards therefore have a very good material bond, so that no larger tolerances for the assembly are required. In a stable delivery panel, rigid-flex printed circuit boards as well as rigid printed circuit boards can be assembled as standard.

SMD assembly - are there any restrictions regarding min/max sizes for the components with rigid-flex?

No, rigid-flex printed circuit boards are made using the same components as purely rigid printed circuit boards. For the routing of complex components such as UFBGA with pitch 0.5mm or even 0.4mm, adapted HDI technologies are necessary and possible. If you have any questions, please contact our hotline at flex@we-online.de.

Detaching the boards from the panel, how is it done and does it limit the positioning of the components somehow?

Basically the same processes are used as for rigid printed circuit boards. One difference is that polyimide foils are tough and cannot be broken like rigid FR4, but only cut or punched. This must be taken into account when designing and separating bridges in the flex area. A good alternative is laser cutting with micro bridges in combination with lift-off, which can be separated manually without tools and without damaging the flex area.

Attention: Never separate webs with side cutters! This will damage the PCB!


How much does FR4 Semiflex cost?

FR4 Semiflex is based on rigid circuit boards with optimized stack-ups. Additional expenditure is needed for z-axis controlled depth milling and flexible soldermask application by means of screen printing.

Compared to heterogeneous systems using cables and connectors there is a cost advantage at least on system level – for shielded cables and connectors even on component level with better performance, miniaturization and reliability at the same time!

We have analysed this on the basis of a real example in our webinar.

What is the price difference of a FR4 Semiflex with 2 layers and 1 bending area compared with a standard 2 layers pcb?

The price of a FR4 Semiflex pcb includes that of the rigid board plus the surcharge for the processes for presicion depth routing and application of the flexible solder mask. As the expenditure for the routing is layout related the additional costs for this process could be in between 10 and 100%. We offer price estimation quotes for your design based on a drawing or sketch.

Things to consider in terms of panelization?

We are happy to support you in this. The aim is an optimal compromise between stability and separability. The optimal design of the delivery benefit often also has a great influence on the material utilization and thus on the price! If you have any questions, please contact our hotline at flex@we-online.de.

UL label

Is there a UL-Listing available for flex-rigid?

Yes! Würth Elektronik has several types UL listed. All types have the flammability class according to UL94 (V-0 or V-1), some types have a so-called "full recognition" and are additionally listed with the parameters MOT, CTI and DSR marking.

The requirements regarding UL are derived from the UL standards for the UL category applicable to the product (Guidelines, UL certification documentation: "Section General" or as "critical components" in the "Description").

For many applications a flammability class V-1 according to UL94 is sufficient.

Which possibilities are there for UL marking?

There are three mainly used possibilities for UL-marking on PCBs:

  • Add the marking to the pcb-layout in copper on top- or bottom-side
  • Place the marking within the soldermask on top- or bottom-side (not over copper filled areas!)
  • Print the marking over soldermask by silkscreen printing.

The most keen options are Versions 1 and 2. Variant 3 is the most expensive version if no service-print else would be required on the board. But position can be choosen almost everywhere (except areas where soldering is required).

Which size for UL-marking is needed, which content?

  • Digits/letters should have in minimum about 4mm in height.
  • There is minimum space required for our company marking ´WE´, followed by a two-digit-number describing our board-type and additionally a circle for the marking of the production plant.

Example: So this would look like “”WE50 O”, where the optional mirrored ´UR´ is already economized. There is no regulation to mark the flammability rating like “UL94 V-1”.

If marking for Canada is required (ZPMV8), then the addition of the ´cURus´ is always necessary.