USB VID/PID numbers for small manufacturers and such like

[yourwebsiteisbr]

I work for a small company and have driven the development of a new product, with MK20DX256VLH7, which uses simple Serial over USB to communicate with management software on a PC. For this one product it is far from worth our while to get our own Vendor ID: USB.org - Getting a Vendor ID

Recently I was approached by somebody with a device which reports as a generic keyboard HID using a bogus VID (simply unassigned/unknown) where the problem was that on MS and (proper) Linux machines this device was accepted and input from it was immediately recognised but when it was connected to an Apple machine running OS X there was a stuff around for the end user to be able to utilise the input and the only thing I could eventually pin it on was the bogus VID/PID combo. (Small Chinese company was their vendor, no doubt similar boat to the one I just barely float in...)

Basic upshot is that it occurred to me to request the manufacturer of the processor in use for that other device to publish some generic HID product IDs (PIDs) under their Vendor ID for their customers to use without restriction so that small companies, and hobbyists trying to go commercial, can use them without fear of retribution or pedantic behaviour such as Apple appears to be showing there.

Freescale Vendor ID is 15a2 and NXP Vendor ID is 1fc9 (formerly Philips VID=0471) - my request to you guys is to publish generalised product IDs under each of these VIDs; for my purposes "Generic HID" and "Generic Serial/Com" will suffice.

I think that the benefits these, and other PIDs others will probably suggest (given the idea) will give your customers should get your company a lot of kudos from various communities of users of your microprocessors.

[Hui_Ma]

Hi

Freescale(now NXP) is able to provide free VID/PID to customers who have small volume production (less than 10k). Customers can use the Freescale(now NXP) VID with an officially allocated PID. Please contact with your local distributor to get help to apply VID/PID.

You can find your local Freescale(now NXP) distributor from below link:

http://www.nxp.com/about/worldwide-locations/worldwide-sales-offices/distributor-network:DISTRIBUTOR...

Wish it helps
Have a great day,
Ma Hui

-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------

[yourwebsiteisbr]

Thanks Hui_Ma, If the devices I am talking about required specialised drivers then this would be worthwhile BUT I am talking about all the small things that well known (and already included in Windows, Linux and [even] Mac OS X) drivers support; generic keyboard, generic mouse, generic joystick, generic serial via USB etc etc.

For example, I saw this earlier in another discussion: Re: Windows 8 USB CDC Setup (or .inf) file ?

"If you want your CDC device to work in any Windows 8 Machine (32 and 64) without need of drivers or .inf files you can set your stack VID to 04D8 and PID to 000A and your device will be recognized as Serial Port automatically, windows 8 has built-in drivers (which is in reality the same drivers used by the inf files posted in this thread that are made by Microsoft) associated to this VID/PID pair. The downside is that in the Hardware manager the manufacturer will show as "Microchip Technology, Inc.". I know it's a nuisance to show a Freescale device as a Microchip but somehow Microchip managed to make Microsoft ship a .inf file with windows 8 to recognize their devices automatically and other manufacturers didn't"

The amount of times I have plugged devices into Linux systems and issued 'lsusb' in the terminal to find that the device is using an unassigned VID, or a VID which clearly does not have anything to do with any 'badge' on the given device, is ridiculous - the keyboard I am tapping this in with at the moment is a prime example of this; the Vendor ID does not match the badge and the product ID is not listed under that Vendor ID in the list (hopefully being) maintained at http://www.linux-usb.org/usb.ids

That's great that we can (basically) apply for a PID to use for our (actually) unique device but for something that is just going to use a simple USB interface to provide some or other functionality [a USB stick which pretends to be a keyboard and sends a random keypress from a predefined list of keys at a predefined interval so that a player does not get timed out on a server for an example] an actual unique identifier is somewhat ridiculous.

I honestly think that if NXP publish generic device PIDs for small companies, and hobbyists (members of the maker communities all over the world), to use for the 'well known' interfaces then it will be very good for marketing your processors. - generic mass storage would be a good addition to the list too.

[mjbcswitzerland]

Hi

A generic PID would not be in accordance with usb.org since a PID refers to a particular product.

In cases where driver customisation is required potentially having two products being recognised with the same PID would possibly stop one of them working since the host wouldn't be able to distinguish one from the other.

Freescale/NXP offers a simple solution for a dedicated PID for small(ish) product quantities, which I and various collegues have made use of during the last 7 years to simply and cleanly resolve the issue.

I think that the issue here is more usb.org. The fact that semiconductor manufacturers are allowed to issue PIDs for their VIDs is already an exception. VID owners usually are not allowed to do this so we should be thankful of the exception since it is saving us $5k a year and just requires filling out a simple form.

If NXP were to do something in the "gray zone" it may rock the boat and end up with usb.org even rethinking its exception policy and then everyone would much worse off.

Regards

Mark

[bobpaddock]

It is only a yearly fee if want to use the USB logo.

If don't care about the logo it is a one time fee.

Yes, the fee is far to high either way...

[mjbcswitzerland]

Bob

When I started with USB (7 years ago) a VID cost $2k. Now it costs $5k.

I am glad that we have had low inflation the last few years otherwise it would probably be a lot higher...;-/

Back then there were in fact a number of companies with a VID that would sell PIDs to you for a few dollars. The terms of VID ownership were not that clear at the time but these companies soon stopped these offers after getting contacted by usb.org's lawyers. The semiconductor manufacturer's exclusion from the ban to allocate PIDs to their silicon users still surprises me (that is, that it is still allowed) but it does make things much easier for first-time USB manufacturers. Imagine having to explain to the project manager that the $0.5 USB interface in the new gadget can't be delivered until a $5k payment is made. Many project managers will not have planned or budged for this (since they didn't realise it before) and it can be a horrible shock in the final days of an already over-budget product development.

Lets hope that the availability of NXP PIDs continues in the future. If the USB products are successful, increase beyond 10k pieces and generate enough profit to afford a VID then fine - it is the potential hurdle at the beginning that needs to be avoided.

Regards

Mark

[yourwebsiteisbr]

FYI Mark: Nobody so far has posted anything to this discussion that I did not know before I made my opening post.

The processor manufacturer that comes to the party on this request of mine will probably garner a great deal of the maker (and small business, which just want to use simple well known and completely driver ALREADY supported interfaces) markets out there.

[mjbcswitzerland]

Robert

It can't be known what you did or didn't know before but we know now that all contributors are at the same level of knowledge (and forums are used as reference by many, so others can also synchronise to the same level of knowledge - or perhaps add more contribution to push it up more).

If you are right, NXP will take up your advice. If they don't it doesn't mean that you are not right but more likely that they don't actually care much about makers or the little businesses - now if it were a way to attract automative then I am sure they would prick up their ears though ;-)

My experience, after working on several dozen USB projects (mostly for smaller businesses), is that they found working with Cypress, Microchip or FTDI in the past easier (for whatever reasons). But I don't remember a case where applying for a PID actually caused any impact on decisions - in general the silicon choice and design had aready been mostly completed before such questions were even considered. Also USB logos and testing are generally of no concern (unless your company is a house-hold name like Logitec or such), also not for big industrial concerns. They prefer not having to deliver drivers and will organise certified ones if needed (to avoid the user having to force the istallation of what otherwie gives the impression it could be a virus infected process) but want the operation to be robust, compatible and efficient but not needing to have a logo on a box that is essentially saying that it will be so.

Regards

Mark

[yourwebsiteisbr]

For a non-profit they sure seem cash hungry, those usb.org people.

####:#### <Processor 'X'> as HID

####:#### <Processor 'X'> as CDC

####:#### <Processor 'X'> as Mass Storage

Basic point - it is 'processor x' running the USB stack providing whichever interface, right?

If every single little keyboard, mouse, joystick etc etc got its very own PID - just seems ridiculous, sorry.

[mjbcswitzerland]

Robetr

It is per "product" - there is then a serial number in the device descriptor that can be used to identify single devlces.

Regards

Mark

[yourwebsiteisbr]

you miss my point entirely Mark - only companies like dell are going to screw around individually serialising every individual keyboard of the one make and model.

How many piddly little companies do you think have made (at least almost) unique keyboards? My guess would be approaching the thousand mark and I am willing to bet that the majority of them just copied someone else's VID/PID values or just used completely bogus ones.

I am contemplating making a joystick which emulates a keyboard, mouse and joystick - it is just HID imho.

[mjbcswitzerland]

Robert

I expect there are many companies who don't respect rules (recycling MAC addresses or using bogus VID/PIDs, for example, or forging safety or EMC declarations). Some of the rules are gentlemen's agreements and some are law (with risk of being fined, imprisoned or banned from a profession for a period of time).

Individuals in the maker's community can use bogus everything in their "own" environment (if there is a collision they know why and can fix it and no one outside of their environment cares or knowns about it - I am sure that usb.org will not file a lawsuit against them even if they were aware of such a case).

As soon as a "product" is in question which is being sold to consumers and takes part in the global market place (takes part in consumer PC environments / Internet) I simply would not take any chances.

Most Kinetis parts have a built-in 128 bit unique ID which can be used to serialise devices and in my experience manufacturers tend to want to be able to trace every device that the have sold (even smallest companies).

Regards

Mark