Sensor Modules Help Accelerate Embedded Vision Development

Sensor Modules Help Accelerate Embedded Vision Development

Sensor Modules are small imaging solutions used to easily integrate customized vision technologies
into machines and devices that allow them to see.

 

Devices and machines, equipped with embedded vision technology, are mostly custom solutions. Often, they are made by contract manufacturers or OEMs and play a crucial role in keeping development cycles short while getting to market quickly. Releasing weeks in advance can make the difference in winning a market or getting a good head start on competition. Imaging solutions made with small, compact and purpose-built sensor modules can provide this advantage. For these sensor modules, pre-existing sensors are placed on a PCB with a standardized connector with enough circuit conditioning to get the sensor up and running quickly. In many cases they come equipped with a lens mount and an adjustable lens. Often used for consumer applications, the sensor modules make it easy to include vision technologies into devices. Ready-to-integrate sensor modules, with their small footprint, provide raw imaging capture and are directly connected to ISPs, GPU’s or natively integrated processors like Nvidia, Jetson, Raspberry or Qualcomm. For vision engineers and developers, these modules speed time-to-market and optimize resources from prototyping, to testing and mass production.

FRAMOS’ Ecosystem of Sensor Modules and Adapters

FRAMOS, the imaging specialist, has professionalized this modular approach. In addition to pre-soldered sensor modules, the FRAMOS Embedded Vision ecosystem provides standardized connectors from sensor module to adaptor board and from adaptor board to processing board. With these mating connectors, processor boards can easily be created to support any sensor module. André Brela, Product Manager at FRAMOS, says:

“The ecosystem is built on a very flexible platform that provides the capabilities and functionalities to evaluate not only the sensor but how it performs on various processor development boards. Even if a customer has to change the processor in the middle of his process, this only effects in an additional adapter. All the evaluation and testing efforts remain untouched, as the sensor module itself stays as is. Every component is interchangeable making them very flexible during the design phases and provide an excellent reference for the final production implementation.”

According to Brela, these products provide the tools needed for vision developers and engineers to quickly get to a working proof of concept, and then evolve it with the provided schematics and board design files into their end-product configuration.

FRAMOS-Embedded-Vision-EcosystemPicture 1: FRAMOS’ ecosystem of interchangeable sensor modules and adapters


Accelerating Product Development Process

Phase 1) Proof of Concept

The product development process mainly consists of five phases. Phase one is the Proof of Concept where an idea is tested to see if it is valid for the end product. At this point, engineers need to make sure that the choice of sensor and processor are sound for the end system. They are looking to ensure that they can get the right image data out of the sensor, at the speeds needed and that the processor can process this data and provide the necessary results to fulfill the application requirements. Using sensor modules during this phase gives engineers the ability to select from a family of sensors and processing platforms to find a quick solution without any compromises. It is “like buying a kit and get it running”, as Chris Baldwin, Technical Service Manager at FRAMOS, says. “Engineers can buy five modules, a couple of adapter boards and one processor to see which combination fits best”, he continues. “Not only can they choose from different options, they can use different sensors and processing boards and have them all running, all at the same time.” Basically, it is possible to develop two systems with different processing levels in parallel on the same basic platform. Sensor modules can be used as sensor evaluation kits as well, but better, as they can more easily have them running more closely to what the end-product will be. Software developers can work with a full kit to start development on their application while the hardware team is working on the final hardware design, in parallel.

“Besides the pure sensor and adapter choice for the embedded vision application, many customers not only want to avoid effort, but they may also lack experience in designing sensor-based hardware and software. High-speed signaling and the design of a sensor board needs deep knowledge and engineering skills, plus a significant amount of time and resources. With existing sensor modules, there is no need to spend time designing and prototyping your own sensor board,” adds André Brela.

From a technical perspective, engineers can leverage the ecosystem to get access to the full stream of raw image data and are able to use the whole system at full speed, not just the image sensor. Classic EVB’s do not have the bandwidth to get the full data rates of the sensor. This makes a sensor module better for validation and testing. Engineers can evaluate multiple SKU’s, multiple price points and options, in a controlled and familiar development environment, while getting images in minutes. They can benefit from open source software that is Linux based and supports Open CV and are backed by a great support community. Starting with sound, known good, hardware is key to making the proof of concept simple, quick and successful.

 

Phase 2) Prototyping

During the second phase, the Prototyping phase, sensor modules create a test bench from the evaluation platform, which assist in debugging new hardware and software by providing a working reference. Prototypes are the first samples of the final product, so they need to have all the functionality needed, including all the ISPs/FPGAs and/or microcontrollers running the final software. For the engineers, the modules are an excellent example of how to create your own hardware, interface to the sensor and ensure that all the hardware can be tested easily.

“Known good hardware is a theme for the whole ecosystem”, says Baldwin. “For building something new, to extend features, or to simply try new things, proven hardware is the basis to ensuring a successful end. Engineers need to have confidence that their reference is bug free when they are chasing bugs in their designs.”

As the sensor modules come with proven hardware, the question, “Is it a hardware or software bug?”, can easily be answered with their designs. The engineers can rely on sound components to track down the source of their issues.

Trying different sensors and processing board combinations is as easy as clicking boards together. Therefore, the sensor choice will not dominate the development, while bringing with it freedom and flexibility. The modular approach used in the Ecosystem is like playing with LEGO bricks and allows for homogenous development. It is not necessary to build multiple prototype boards to ensure that there is enough available to validate the design. Leveraging a preexisting design for the prototypes and the end-product remove the need for unnecessary respins to correct mistakes in the design of the original component.

 

Phase 3) Pilot Production

Once prototyping has been finalized successfully, phase three, the Pilot Production phase, starts. In this phase, sensor modules again provide the ability to rely on standard modular hardware. The first production run is normally the most worrisome one. Using sensor modules, manufacturing engineers can ensure that there are no Contract Manufacturing (CM) production issues with the first batches of products, which effectively leads to 100% yield. For a company, this allows for quick reaction to sudden sales demands or to quickly create different models for different target groups.

“It’s the 120-pin connector that easily allows the building of different versions of a product,” as Technical Service Manager Baldwin states, “the sensor can be swapped in and out as needed.”

Having hardware that easily connects together makes the manufacturing process simpler and more reliable. Using off-the-shelf products means that inventory levels can be kept lower, which is especially important when products have multiple build options, using different sensors and sensor technologies, and very long lead times. Incoming testing and QC can also leverage parts of the Ecosystem to test newly build components from CMs to confirm their operation and adherence to specifications. Again, in this phase, production test suites can be created with the Ecosystem components while the first run designs are being built at the CM, further improving go-to-market timelines.

 

Phase 4) Mass Production

Mass Production is phase four and it benefits from sensor modules for product BOM costs and cash flow. For smaller volume productions (<10,000 units), the reasonably priced modules leverage the overall buying power of one CM working with multiple customers, while also saving costs through easy product manufacturing processes and simplified incoming inspections. Like with any company but more importantly for smaller ones, cash flow is always a big concern. Using sensor modules that are already on the shelf at a supplier minimizes the need to carry inventory to cover component lead times while providing protection from excess inventory should the new product have a longer adoption cycle in the marketplace.

This is the last point in which one wants to have a quality issue. Companies transitioning from small volume manufacturing to high volume often find unforeseen issues creep up as they move from one CM to another, one that might be in a foreign country with less skilled labor. Therefore, using modules provides a robust and low-cost approach to mitigate these issues while scaling BOM costs with volumes while minimizing the impacts of changes within the supply chain.

 

Phase 5) Next Generation

The benefits of sensor modules do not stop once the product is developed and selling in the market. Phase five is about innovation and describes a fast way to enable the Next Generation of the products. With a strong understanding of the Ecosystem in place, engineers can look at new components like different sensors or processors to further extend the existing products. They can quickly add in the new component into the existing product that readily support it and see how it performs. Their familiarity with the Ecosystem makes the PoC quick and thus allows for quick design and development of this next gen product further pushing out the competition as this new product comes to market quicker than before. The modular design allows incremental or large steps in technology to be incorporated because they leverage the basic platform, the Ecosystem that allows simple interchanging of sensors and processors. This approach allows for simpler and lower-risk R&D choices when implementing innovation into new products.

 

Features to Consider When Selecting the Right Module

The question of which module to choose always includes matching the sensor characteristics with the application needs. Information like optical format, resolution, shutter type, and pixel size are all key and engineers can choose from a broad range of available sensor modules with sensors from various manufacturers. Also, sensor interface specifics like clock frequency, voltage and data format, and mechanical and environmental specs like lens mount, size, operational temperature and humidity values help to select the right model that maximize the performance of the imaging system.

Also, having working software drivers for the modules on the processing boards greatly reduce the time needed to develop their products. Software engineers need to know how these drivers are built, what OSes they support and what additional processing libraries are available to further extend their processing power in their final system.

Mechanical drawings of sensor modules and other adaptor boards allows for quick and easy integration into CAD programs. A mechanical engineer can then design how they will be incorporated into the final design by factoring in the size and placement of the board and sensor while ensuring the optics are all aligned to the optical center of the sensor. Datasheets are a hardware engineer’s main source of information for designing in these components. They explain the connector pinouts and provide electrical details of each module that enable them to be utilized into a new design. On top of this, schematics, gerbers and layout files are provided to allow these engineers to redesign and more closely integrate them into their end designs while removing unneeded portions of them to minimize their footprint and reduce their costs.

 


Info-Box: Key to the overall flexibility of the FRAMOS sensor modules ecosystem is a 120 pin-connector. This 120 pin-connector is the abstraction point where sensors and processors connect. It is the passthrough for power, data, commands and control signals. Using these adapters, customers benefit from this “mix and match” and “plug and play” approach that the Ecosystem provides allowing for more flexibility in changing components at every stage of the development process.


 

The full FRAMOS ecosystem is extremely helpful in saving time and resources during the evaluation and prototyping phases. As the first production systems are build, issues can be more easily tracked down by using the Ecosystem components as references to a known working system. These same components can then be adapted to be part of the test fixtures for new products being built at CMs or to validate the assembly of component on the production floor.

Eliminating Business and Development Risks

Image Sensor Modules eliminate design, business and time risks typically tied with the integration of imaging into machines and devices. Especially for small and medium-sized companies, and start-ups, sensor modules offer a fast way for converting ideas into applications, allowing one to start testing the final concepts much earlier in the design phase. Engineers can see, at a glance, how the system will work, allowing for the adoption and modifications of settings, or to try new sensors or processors. From start-ups to high-volume OEMs, scalability is a main advantage of sensor modules. The ecosystem enables customers to use the same off-the-shelf sensor module in the early product development phases, while allowing to be used effectively in high volumes as well due to its cost optimizations as quantities increase.

Stay Tuned with FRAMOS. Sign Up for our Monthly Newsletters

Contact Us

Salutation*
First Name*
Last Name*
Job Title*
Company*
Email*
Phone*
Country*
What FRAMOS product or service would you like to know more about?*
By submitting the request, your data will be processed to get in contact with you and answer your questions. For more information about the processing of your data, please take a look at our  privacy policy.

Contact Us

Salutation*
First Name*
Last Name*
Job Title*
Company*
Email*
Phone*
Country*
By submitting the request, your data will be processed to get in contact with you and answer your questions. For more information about the processing of your data, please take a look at our  privacy policy.