Designing RF generators for MedTech applications
Add in the continual need for the prolonged usable lifetime of a product and the designer will have to ensure that all eventualities are covered before moving into production. This article should give anyone involved some idea of the key hurdles and some ways to overcome them.
Customisation of magnetics
An initial decision is whether custom magnetics are needed and, if so, what will be required to meet the challenge their usage may bring. Some applications will use custom magnetics because there are no alternatives and there are off-the-shelf magnetics for run-of-the-mill applications, or pre-made DC-DC converters. The drop-in ones may prove best in some cases and will save engineering time for more complex issues with critical therapeutic parts of a medical device design.
However, where energy is crossing the patient isolation barrier, operating at a specific frequency and with tight performance constraints, no off the shelf product will get near to meeting the applications requirements.
Custom magnetics are ideal but in RF generators they can introduce extra challenges such as creepage and clearance distances, EMC issues, leakage current and low capacitance requirements. Without the competency to develop individual components to meet these challenges it becomes necessary for an OEM to get a custom magnetic manufacturer to design for a specific requirement. The problem is, the vendor then owns that IP and the OEM may well not get access to the full design information. When the need arises to source parts in a different region, to achieve more competitive pricing or an increase in volume above the original supplier’s capabilities, it could be a problem. Ownership of the design gives the ability to take that part to a custom magnetic manufacturer in any location.
Working with an EMS on the project means the product developer should have a manufacturer with an understanding on how to achieve true manufacturability of the custom magnetic component. There are will be performance and cost trade-offs to make during the design process so a good working partnership is essential. For example, using a more exotic material may be better for performance but increased cost and longer lead times may well counteract that incremental gain.
Relating all the information that a designer has to a magnetics house effectively can be difficult too and, if that is achieved, it is still in their hands to make the right choice for your product. It is essential, therefore, to only work with a custom magnetic manufacturer that is the right fit for the design and that gives guaranteed access to the IP for global use.
Medical process capability
When developing RF/piezoelectric surgical generators, which are classified as FDA Major Level of Concern and IEC62304 Class C. it is essential that the medical regulatory process to be followed is understood, including IEC60601-1, IEC62304 and ISO14971. When the CE mark is required for the medical device and the medical device’s software plays a role in the patient or operator safety of the device, IEC62304 is applicable and the IEC62304 TRF (Test Report Form) must be filled in to demonstrate compliance.
With the release of IEC60601-1 3rd edition, the definition of essential performance was defined as performance of a clinical function where loss or degradation beyond the limits specified by the manufacturer results in an unacceptable risk. Furthermore, essential performance must be maintained through single fault conditions. A surgical generator design team must therefore understand the implications of this requirement on the architecture of the system to allow essential performance to be met against this strict definition. Utilisation of redundancy (sensing circuitry, microprocessor controls & supervision, etc.) is almost certainly required to ensure this compliance.
Helpful DSP controls
RF and Piezoelectric surgical generators typically require ability to sense output voltage, current, impedance magnitude, impedance phase and real power. In order to limit the usage of feature specific ICs that can create parts availability issues in the coming years, and to provide flexibility in the capabilities being provided, it is often beneficial to architect surgical generator controls and monitoring using Digital Signal Processors (DSPs). A DSP with on-chip high speed ADC inputs and PWM outputs is capable of providing the monitoring and control needed for these applications.
Employing monitoring schemes that use dual sample and hold ADCs to allow simultaneous sampling of voltage and current, each signal is oversampled to allow reconstruction of each waveform with enough resolution to provide an accurate measurement of voltage and current. Point by point multiplication of the simultaneously sampled voltage and current signals, yields a real power waveform. From that, an average real power value can be easily computed.
Capacitors also frequently require customisation in medical applications. For some seemingly simple applications it would seem that any compatible part should do, such as blocking caps for RF generator. You could think that with the need for a capacitor of a certain value and voltage rating you can just drop one in. An understanding of the design standards for failure modes (60601-2-2 specific standard) is needed to know whether redundancy is required and what will be hazardous to the patient. It would compromise the production timeline if at the UL or TUV stage the designs are returned to the drawing board because of a wrong component selection.
There is, therefore, no such thing as “just a cap” in medical applications. With many capacitors such as decoupling caps and ‘popcorn parts’ in a design that have nothing special about them, the skill is to know where the critical parts are and why they are critical. An established team with overlapping circles of design responsibility will be able to define, even early on in the architecture level, details such as how to achieve high-bandwidth data across a patient isolation barrier.
Designing onto the PCB
The PCB design is critical to the usability of any medical device, with the need to have an understanding of what the physical design of the board means for the broader restrictions of safety testing, creepage and clearance, spill tests, moisture ingress and touch temperature. All these pose challenges. If the placing or mounting is incorrect, whether on single or multiple boards, the likelihood is that 95% of the requirements will be met but the other 5% will require a redesign of the whole board.
Engineers understand what has to happen with the PCB, but let them do the layout and the board may well prove non-manufacturable. However, the PCB designer must understand the requirements of the engineer. This means close collaboration is required between design and engineering teams with both bringing what is most important to the design to the table to achieve the best results.
It is the software engineers who have to contend with recent statistics that indicated that around 80% of features in software products are never or very lightly used. It could be said that such features should never have been implemented in the first place.
With a customer that has a grand vision for their product, and a tight timeline, the software engineer will have to persuade them to prioritise the must-have from the nice-to-have features. Omitting features should never be considered optimal, but realistic decisions have to be made. There are many extra features, however, that have been customer defined for forward-looking requirements that may not be utilised in a first or second release.
With future proofing important in fast changing medical device market, upgrades must be efficient and simple to achieve with some purely software modifications and some hardware changes. At an early stage in the design process engineers will define just what hardware needs to be available for these upgrades, even if not initially utilised, and get that in situ to pass verification and manufacturing tests. The future software updates will not require hardware changes because that is in place and tested to meet these new requirements.
An upgrade delivered by USB
Utilising the ubiquitous USB connections it is possible to plug in a USB flash drive with a software upgrade image and have a bootloader or software upgrade management software recognise that image, verify it and install it. It can be complex, however, as it may be more than a single upgrade image and include UI software, FPGA and control software. It is possible to make this a simple, clean upgrade process, which could easily be manages by sales staff in the field, if these are extracted and a protocol put in place to transmit the images from a processor with USB access to the other programmable devices.
The customer will have high expectations but their main priority is the end product that will be presented to the patient, its therapeutic benefit and its user interface. They will not necessarily be concerned about the different elements that have had to be achieved as long as the product meets the original specifications, proves reliable and is delivered on time at the agreed price. By managing customer expectations, and ensuring it has the best design and engineering team, in place the EMS will deliver a medical device that meets therapeutic requirements and with a long shelf life.
About the author:
Anthony Green is Director of Engineering EMEA at Plexus – www.plexus.com