Sourcing of electronic components is a huge market, every electronic device requires multiple types of capacitors, resistors and diodes. Procurement process when generating a bill of material (BOM) involves thousands of large global enterprises (Murata, ROHM, STMicro etc.). This is why it is difficult to imagine why this is the only line of business that people still look at the datasheet ( A user manual provided by the part manufacturer).
With everything going online, procurement has done so many years ago. The idea was that consolidating procurement can save a lot of money for manufacturing and industrial companies but electronics sourcing is upside on his head. First we have the systems but the data is not ready to be processed.
Electronics official distributors hold the data, they present it and receive it from their suppliers regularly. Distributors like any other business are bound to their clients and want to promote certain suppliers according to changing relationships, available inventories, therefore having an objective data based platform is not top priority. Especially that distributors have to manage logistics and inventories, a difficult complex task by itself.
This is why sourcing of passive / active components and semiconductors are lagging behind the ecommerce world and are perhaps the last to start climbing the innovation wave.
Good and uniformed data source is key to digitizing sourcing of electronic components.
But like any other market issue, it has a business side but also a data technical side. Without a good readable and comparable source of data, that is stored in a tabular format, there is no way to create good search capabilities.
Electronic component data is provided from many different sources (part manufacturers such as Vishay, Texas instrument and Analog devices) is a non structured data.
Furthermore, this data is 3 dimensional and changes according to temperature or frequency of required operating points. As of today each manufacturer looks to optimize his performance and chooses measurements to be done at his preferred point, this is sometime different to his competitor optimised point. Intentional or not we don’t know but it makes this data non-comparable and hard to make sense outside the specific company domain. Meaning you could compare murata inductor to another murata inductor but cross brand to a TDK coil is very hard and for obvious reasons.
While there have been few initiatives to uniform electronic data interchange (EDI) it is also at the hand of part manufacturers.
Where it is visible is in the CAD software world, where symbol and 3d models are consumed, in order to allow engineers choose the right part those models are uniformed and now available in several web tools for free ofcoruse. (SnapEDA, Ultralibrarian). There are spice models and s2p (S-param file extension) that can assist in some of the required tasks but not in really comparing those models.
Some companies try to make an effort to normalize the part technical specifications, perhaps the most dominant is sourcingBot.com, which managed to normalize part specifications and is the only company now providing capability for free online.
SourcingBot deployed machine learning and statistical analysis to different series of electronic components.
Ran Oren, CEO of sourcingBot tells us: “We came from the data world, looking at it as a data science problem and not an engineering problem. We were able to deploy normalization processes that could normalize and uniform technical specifications. Unlike many other data science problems, we had an abundance of data to use a training sets for our models.””
This is why sourcingBot is delivering a unique deep parametric search on its platform, as well as at operating point comparison. Focusing on passives and discrete semiconductors, its comparison mechanism resembles what you would expect in e-commerce and automates the expertise required until today.
Ran adds, “It was a tough task, we started with a very wide view and quickly focused on the passives and discretes as they are the main market that requires both cross referencing and an expert system that replaces knowledge that’s just disappearing from the world and is very costly to maintain.
There is no doubt the train has left the station and this world is starting a long climb up the innovation scale, in the way it needs good startups and a step in from different regulators to make a big substantial change.