Selecting the right memory product for the job sometimes means choosing specialist OEM memory. Nexus GB’s John Barrett explains why…
Consumer memory devices, such as SD cards and USB flash drives have some obvious advantages, but for embedded memory applications they also have some downsides: lack of reliability, ruggedness and security.
Q) What is the difference between specialist OEM and consumer memory?
A) Consumer products such as USBs and camera cards are great for storing digital photographs, Powerpoint presentations and so on, but if you want something that you know will still be around in 10 to 15 years time, you would be better off buying specialist OEM memory, such as that available from Nexus.
For example, we have a large railway signalling technology project with a UK rail firm. Right from the outset, they made it clear that once memory technology is engineered in to their design, it will need to stay there for 25 years, at least. That means we will have to keep supplying the same memory product, or an exact equivalent that is identical in function and form, for that period. Most of our customers have similar requirements with regard to obsolescence. Another example would be defence and government applications, such as secure phone lines and crypto ignition keys, where design specifications need to be in place for decades, not years.
Q) Are we talking about detachable external memory?
A) Correct. Datakey Electronics (DKE), the US OEM for which Nexus acts as an exclusive distributor, has a number of form factors available. The main one looks like a key that is inserted into a reader. It makes contact when turned, which means its ideal for cashless vending by manufacturers like Klix who we have supplied with almost three million keys. The other main form factors are a token, which is not unlike a USB in size, and a plug.
Q) What role does obsolescence play in selecting memory?
A) It’s fundamental. There is a story circulating in the USA about a company that had spent a vast amount on R&D to engineer an SD card into a product. They had just launched their new device and a week later they got a ‘last time buy’ letter from the memory supplier. They had to take the memory out and re-design the product, which cost them a small fortune.
In contrast, we first supplied our largest customer in 1989 and we are still able to supply exactly the same product today. This is crucial because, manufacturers of secure phones, railway signalling units and even coffee machines, don’t want to go back to the drawing board every four or five years.
Q) How do you incorporate new, larger capacity memory into existing key formats?
A) When larger capacity memory is introduced with a different pin-out format, Datakey Electronics matches that format to an existing key. Furthermore, we are happy to deal with a range of product quantities from tens or even hundreds to hundreds of thousands. In contrast, consumer memory manufacturers are only interested in quantities of several millions.
Q) Are there single sourcing risks associated with OEM memory?
A) Given that genuine OEM memory can only be sourced from Nexus in the UK, and Datakey’s other distributors across the rest of the globe, purchasers could be forgiven for thinking this is an issue. Datakey Electronics however, ensures that it doesn’t source the memory chips that it over moulds into keys and tokens from a single supplier. Instead it qualifies a number of different providers, which secures the memory component of the device for decades.
Furthermore, Datakey Electronics supplies the whole item – not just the token or key but also the connector or reader. That means the customer doesn’t have to search for connectors to match the required pin-out, which they would have to do if they were using SD cards. If something goes wrong, the connector manufacturers will blame the memory manufacturer and vice versa. With Datakey Electronics supplying both, if there is a problem, we provide the solution. This secures the entire item, not just the memory component.
A) Specialist memory is more expensive initially for the reasons outlined above. In the longer term however, it is cheaper because you don’t have to replace it as frequently. Engineers tend to look at long term return and it makes sense for purchasing professionals to match the designer’s viewpoint.
Furthermore, the casing that contains the board and its components is often the most expensive part of a memory device, particularly if it uses special materials, but also in terms of R&D. If the connector has to be redesigned after a few years, that casing may have to be changed at great expense. With specialist OEM memory, this should never happen.
Q) Can the cost be reduced by using a lower capacity memory product?
A) Most embedded applications require kilobits or megabits, not gigabytes or terabytes, so for applications like access control, parameter upload and firmware updates, a specialist OEM EEPROM or serial flash based data carrier system has a similar overall cost, but is more rugged, more reliable and comes with long-term availability.
If you only need to store a password on the device, a memory module of a few kilobits would be sufficient. Engineers, however, tend to concentrate on functionality first and cutting costs second, particularly when manufacturing volume is low. Some equipment may even use separate tokens for data logging and security/access control. So, from a purchasing perspective, it’s always worth checking whether the memory capacity recommended, is actually over specified.
For example, an embedded program written in ladder diagram will not require large amounts of memory, particularly if it is encrypted. So, , in these cases costs can be cut by opting for specialist OEM memory and not paying for the extra, unused, storage.
A) Yes, specifying industrial memory makes it possible to control the way customers source replacements. An OEM customer can’t go to Dixons or Comet, or buy the memory product online, providing the OEM with a guaranteed revenue stream as a result.
In fuel dispensing, for example, many companies will be operating their own pumps with attendant control systems that ensure only approved vehicles are filled up. With hundreds of thousands of keys in daily use, it would be a nightmare for the OEM if users could pick up a product from the local electrical retailer.
Q) Do physical specifications, such as IP67, dictate specialist memory?
A) I would be surprised if an engineer could implement autoclave resistance or gamma sterilisation, on consumer memory. Even if they could, it would cost hundreds of thousands in R&D time, making the memory device very costly!
In contrast, we can supply IP65, IP67, autoclave compatible and gamma radiation resistant tokens as well as EMI-enhanced memory. The latter is particularly important in defence applications where hacking may be a threat. Some of our connectors also have metallised EMI shields. The bottom line is that if you want specialist functionality you have to buy OEM memory as consumer products aren’t an option.