Long, narrow flex circuits push manufacturing limits as catheter designs demand more connectivity.

I am asked quite often about the longest and narrowest flex that can be made, especially as new applications emerge that support minimally invasive surgery and pulse field ablation procedures used to treat atrial fibrillation (AFib). I will assume readers are looking for the longest and narrowest flex circuit that can be built for a reasonable price. Almost anything can be done “in a beaker” if cost is no object, but if you are looking for guidance on a flex that can be built in volume and not break the bank, there are limitations.

Typically, when I am asked about length and width on the same circuit, the end-application is usually a catheter. The next question after length and width limits is, “What is the smallest trace and space you can do?” Etched feature size and circuit size are joined at the hip. As circuit size increases, the ability to create very small features at a reasonable yield becomes increasingly difficult.

Figure 1. Example of an ultra-long, narrow flex circuit used in catheter applications, shown with a coin for scale.

Figure 2. Close-up of a miniature catheter flex circuit with fine-pitch features, shown with a coin for scale.

Currently, what I would consider reasonable feature limits that can be built with good yields are:

• 50µm-75µm trace/space (1/3 oz. copper or thinner)
• ~1.5 meters in length
• ~2mm-3mm in width

These limits are regularly violated, however, as the medical industry continually pushes for more connectivity in narrower catheters over longer distances. It usually becomes a negotiation over which features are the most critical. Since a certain amount of connectivity is typically required by the application, and the catheter inside diameter (ID) is usually already set, we typically focus first on the overall length to make the flex affordable. Two-meter length is the common “ask” we receive on most new designs because that allows the most flexibility in an operating room for patient, equipment and surgical staff.

To make a circuit very long (beyond standard panel sizes), much of the processing needs to be completed using roll-to-roll (RTR) processing, which limits the amount of operator contact with the material. A very large portion of yield fallout on these types of circuits is due to damage from handling. Catheter flex circuits are usually made from very thin laminates, and once a panel is longer than an operator’s “wingspan,” it becomes very difficult to handle without some amount of damage. While much of flex circuit manufacturing is compatible with RTR processing, very few manufacturers have all the equipment necessary to do all these processes “touch-free.”

One of the first steps in manufacturing a flex circuit is drilling holes for plated vias (assuming the design is two layers). Most mechanical PCB drills have a work area limited to ~18" width and ~24" in depth. While laying material across multiple drill heads to increase the working area is possible, it can cause a lot of headaches, and I’m speaking from experience. The best solution for long circuits is laser-drilled blind vias. This can be done in an RTR process with the proper equipment.

The next step is copper plating the vias. This can also be performed in RTR fashion on a horizontal plating line, provided it is set up for RTR processing. Obviously, if it’s just a single-layer flex, the two previous operations would not be required.

After the material is plated, it must be coated with photoresist and imaged. Imaging is one of the more challenging steps in fabricating very long circuits. Most imaging equipment is set up for panel-based processing and is very limited in its support for circuits longer than the machine workspace. A few printers will do RTR processing on panels that have been attached end-to-end. These printers use artwork films and will image several panels at a time, then index the material and image several more panels.

But even these printers are not suitable for catheter-type applications because the ultra-small feature size would require virtually perfect registration for each printing cycle. If the features are only 50µm wide and you misregister even 25µm in during printing, big problems occur. The best solution is a laser direct imaging (LDI) printer that has a) both feed and take-up rolls, and b) either an ultra-long imaging workspace, or the capability to perfectly register each print cycle to permit blending the multiple printed images.

Most DES (develop/etch/strip) lines are capable of RTR processing if they are equipped with feed and take-up rolls on each end. The chemistry of each section needs to be monitored and adjusted so that each section can run at the exact same speed.

The material will still be fragile when it exits the resist stripper, but less prone to damage from this point on since the features are already defined.

Cover material or LPI solder mask can both be applied in roll form. Cover openings must be predefined (drilled, punched or lasered). Then those openings must be accurately registered over the very tiny etched features. This can be very difficult on long circuits due to dissimilar stretch/shrink between the etched panels and covers. If the cover is LPI solder mask, the same challenges described here for feature imaging will apply.

The last challenging step is excising the parts from the surrounding material. Flex material is not dimensionally stable, which is where the overall circuit width will cause issues. To achieve maximum connectivity with minimal width, most of the circuit's width is filled with traces, leaving very little room for a healthy border. The ultra small border coupled with dimensionally unstable material rules out the use of a punch and die; also, a die that size would cost more than a space shuttle. This leaves laser cutting as the only viable option to remove the parts from the scrap material. Very few lasers will have sufficient workspace to cut out a 1.5+ meter long circuit, so in most cases this will have to be done in sections with lots of fiducials to keep the laser on target.

It should be clearer now why there is a “negotiation” over circuit features to meet all application requirements while still providing a reasonable cost to the customer. It is highly recommended that you have these conversations with your flex vendor early in the design cycle to ensure that you design a catheter flex that is functional and cost-effective.

Mark Finstad is director of engineering at Flexible Circuit Technologies (flexiblecircuit.com); This email address is being protected from spambots. You need JavaScript enabled to view it.. He and co-“Flexpert” Nick Koop (This email address is being protected from spambots. You need JavaScript enabled to view it.) welcome your suggestions.