Upon seeing the ArcticLink III VX RGB->LVDS demo at the recent Linley conference, a very astute attendee asked me ‘why do some OEMs use bridge chips and some don’t?’
This was a very good question, and one we’ve gotten before. Accordingly, it makes good blog fodder, so here we go…
From the outset, this would seem like the most obvious reason: OEM A has spec’d an end product with a processor from CPU maker B and display manufacturer C. The specific processor only supports MIPI, while the LVDS is supported on the panel in question. Obviously, the OEM has a mismatch, and one that needs to be remedied for the product to work.
This can be solved by (1) choosing different components or (2) modifying the chosen components
(1) Choosing different components
We addressed this a few months ago, so I’ll be brief. Changing the components will often have the side effect of changing the industrial design or user-level specifications. For instance, in the example above, the OEM could source a MIPI panel to address the mismatch. However, they may not be able to find a MIPI panel in the chosen format or resolution. A lower resolution would change the competitiveness of their product, and OEMs are most often unwilling to do that. Moving to a different CPU also can change product spec’s, along with introducing the need for system designers to learn a completely new processor and perhaps even have to re-architect their system from the ground up.
(2) Modifying the chosen components
The OEM can ask the display maker or CPU vendor to modify their product (frankly, it will always be the display makers, and CPU’s require a huge amount of time and money to change). Now, the key word is ‘ask.’ Very few OEMs have the juice to actually force their wishes upon display manufacturers. Here’s why…
N.b. for the rest of this section: if an OEM sells many million of a specific product, or has a business model of only offering a handful of products that sell in extreme high volumes, display makers will customize products for OEMs at the OEMs request, hoping they can make back their investment on volume. These OEMs are few and far between, and can likely be counted on one hand.
Now, for the rest of the OEM world, any change to an existing display results in cost for the manufacturer, and that cost has to be paid by someone, whether it be the OEM or the component manufacturer. Accordingly, every change is a business decision for the display maker – “does it make sense for us to do this at our cost, at the OEM’s cost, or at all?’. To get to the eventual decision, the display maker will examine the specific program that the OEM is requesting the change on, as very likely the change requested is for a specific model of tablet/smartphone, and NOT the OEMs entire line of products. While a top 5 smartphone OEM may sell 75M units annually, that volume is most often spread over 20-30 different models, with each model having individual degrees of success. Further, OEMs often change components for different regions of the world, which could further reduce the addressable program volume. Once you start cutting down the addressable market, the display maker might not be able to justify either the actual cost of making a change themselves. Quite frankly, this happens in most cases, so much so that many non-top 5 OEMs have simply stopped asking.
We’ve addressed actual cost, but what about opportunity costs? The display market is a fast-moving, ever-changing one. Do display makers want their talent working on modifying products already in production for a specific program at a specific OEM, or do they want their talent focused on the “next best thing” that will potentially sell to dozens of programs at multiple OEMs?
For the sake of argument, let’s say that the display maker does come back to the OEM with an affirmative to the possibility of a change, and the cost of that change is agreed upon by the two parties. Now you get into schedule…an LVDS display doesn’t become a MIPI display overnight. The display maker may need to custom-design the driver IC for the panel characteristics. This design will take time—months in some cases. Then the need to production qualify the new display comes into play, and you’re now likely 6 months before the OEM can receive a production-qualified display to work with (assuming you have success with your first shot). 6 months is often the entire design time for some products. Tablets and smartphone display are key to the products success, and without the finished good products to develop on, chances of failure are high.
So, all this being said…
Component makers (CPU, display makers, folks like QuickLogic) do our best to model our product roadmaps to best fit into what we project OEM systems will look like during the production life of our products. However, as much as we try, none of us can 100% accurately predict the future, thus there will continue to be instances of an OEM needing to connect a display and a CPU that don’t share a common interface. QuickLogic’s ArcticLink III VX5 products will bridge those interfaces plus offer VEE and DPO. And, for those OEMs who have matching MIPI components? The ArcticLink III VX3 offers VEE and DPO in their matched MIPI architecture, and the ArcticLink II VX2 will continue to satisfy the VEE/DPO needs to RGB->RGB systems.