FTDI driver kills fake FTDI FT232??

[a210210200]

If you look into pi long enough ( or e for that matter) you will find a section that is every imaginable number. Are they going to ban these 2 numbers?  Legislate it is equal to 3?

Interesting thing about pi = 3 is that it works fine if your doing rough estimation.

Great example of such order of magnitude estimation, https://what-if.xkcd.com/84/

Very useful in sanity checks, early development calculations.

[TheRevva]

<Snipped out a chunk> (There are plenty of numbers to go around)

Careful dude, those are very similar to 'the famous words' from MANY people / institutions...
Afterall who on earth would want more than 640kB of RAM (Supposedly uttered by 'some fool' from Microsoft)
And then there's the whole IPv4 debacle which has a positively HUGE 32bit address space...  Who could EVER have imagined we'd exceed THAT?
You may as well also include the good old bar codes (such as UPC / EAN13).  What single manufacturer would ever exceed 99999 unique products? Is 99999 unique manufacturers sufficient to cover the entire GLOBE???

Correct me if I am wrong, but my understanding that both the VID and PID are constrained to 16 bits each.  (Thus a range of 0-65535 each at best case)
It's only a matter of time...
(Just like it's only a matter of time before the signed 32bit 'time' counter in original *nix systems 'rolls over'...  Jan 19 2038 IIRC Sometime around 03:19)

[miguelvp]

http://www.howstuffworks.com/tech-myths/5-myths-about-bill-gates3.htm

[nctnico]

It doesn't (been there, done that). It just enumerates both devices. Remember each USB port has a unique address within the USB device tree as well.

http://blogs.msdn.com/b/oldnewthing/archive/2004/11/10/255047.aspx
In my experience "Exciting things happened" = BSOD.

[madires]

It doesn't (been there, done that). It just enumerates both devices. Remember each USB port has a unique address within the USB device tree as well.

http://blogs.msdn.com/b/oldnewthing/archive/2004/11/10/255047.aspx

In my experience "Exciting things happened" = BSOD.

Another tidbit: http://rtshiva.com/2009/05/19/usb-specification-and-windows-limitation-on-serial-numbers/ which links to the same blog post above at the bottom. I really like udev on my linux systems which is able to fix nearly all issues with USB IDs. In case of the USB serial converters I've written some custom rules to create device specific links. A buspirate gets a /dev/buspirate, an arduino a /dev/arduino and so on. That's much better than finding out which tty the USB device got.

BTW. it should be possible to write a udev rule to automatically fix broken FTDI clones.

[free_electron]

It is very expensive you need to buy them.

so ? what is 4K for a company that designs chips ? dust under the carpet...

the thing is these silicon squatters try to get rich by piggybacking on the success story of someone else. They sell a part that is similar in form fit and function but not a 100% clone. they sell it at a cheaper price than the real one , cash the cheque and laugh all the way to the bank. Meanwhile the people that did the original hard work and built a success story lose their income.

I Vote to bring back tar, feathers and branding (preferrably on their forehead) for this kind of scum.

[zapta]

...I Vote to bring back tar, feathers and branding (preferrably on their forehead) for this kind of scum.

That's not what FTDI did, they chose to retaliate against the users.

[diegosfb]

Nayone knows if there is any way to know if the driver on my PC is the one that will brick devices?
I would like to know that before testing if my devices still work

Regards

[miguelvp]

Nayone knows if there is any way to know if the driver on my PC is the one that will brick devices?
I would like to know that before testing if my devices still work

Regards

Don't be too afraid, FTDI themselves supply the tools to unbrick your device if it does happen and even if it's a clone.

[4to20Milliamps]

I will be happy to send my usb to serial port adapter that worked just fine on every version of windows up to xp, but won't work on windows 7 or 8, back to ftdi so they can verify if it's a counterfeit.

which by the way I'M CERTAIN IT'S NOT.
 
they bricked their own junk a long time ago, but it's different when they do it isn't it? Besides what do these things cost? 20 cents? what kind of moron counterfeits something where there's no margin for profit?

The only people that could.....are the one's that make the chips for ftdi.......they should have taken better care of their business partners


I'm out.

[TheRevva]

The ACTUAL manufacturing costs are probably WELL below the $0.20 that you mention, but that's not the point.
(If it was, I'd expect to be purchasing an I7-Haswell Intel CPU for only a couple of dollars!)
The 'production cost' of almost ANY given chip is almost incidental once you've built yourself a suitable fab and also spent all the time and effort in the R&D of the chip design.
Looking on Digikey, the current RETAIL price of a (genuine? who knows these days) FT232RL / SSOP28 in one-off quantities is NZ$6.34 which lowers to NZ$4.29958 in 500 quantity.
That's somewhat MORE than (for example) a Silabs CP2102 /QFN-28 at only $NZ3.88 each

The R&D costs between the two chips would, I assume, be in the same ballpark and I'd expect the production costs to be equally in the same ballpark.  Therefore, I would bet it's fairly safe to assume that FTDI (and every other 'middleman' in the supply chain) is making about 100% more profit on a FT232RL than on a CP2102.

It's surmised that at least SOME of the clones are simply a repurposed microcontroller.  (One that already had USB and a UART seems obvious).  I would think I could knock up the firmware to 'emulate' a FTDI in a day or two (given the FTDI datasheet).  (i.e. The R&D costs would be insignificant - near ZERO).  The cheapest Atmel with USB and UART I could find on digikey is NZ$4.30 each.  For a suitable PIC, it'd be NZ$2.11 each.
I'm pretty sure I could 'approach' either of them with a request to supply a million 'pre-programmed' chips in a non-standard case.  (Or just buy a million 'bare die' and have them encapsulated in shenzen etc.)
I'm sure as hell not going to though...  That's why they invented Chinese.  Much like the japs of yesteryear, they're incapable of inventing anything much of significance (well, not since fireworks anyway).  All they can do is COPY stuff very cheaply in the myriad Chinese sweatshops.  Heck, just looking at them they're already almost 'copies' of each other??? <Evil Grins @ the 'stereotype'>
IMNShO, companies in the 21st century should NOT be allowed to 'rest on their laurels'.  I considers patents and copyrights EVIL devices designed to holdback progress and innovation.  I believe it's far better to be out 'inventing a new / novel mouse trap' than to waste time protecting last weeks mouse trap.
The only downside is that the exponential rate of advancement is becoming progressively more difficult for Joe Average to keep up with.  (By the time Windows 11 is released, we'll be required to 'reboot it' every 5 seconds.  Back in the Win 98 days we were allowed to wait 5 mins before it needed a reboot).

[a210210200]

<Snipped out a chunk> (There are plenty of numbers to go around)

Careful dude, those are very similar to 'the famous words' from MANY people / institutions...
Afterall who on earth would want more than 640kB of RAM (Supposedly uttered by 'some fool' from Microsoft)
And then there's the whole IPv4 debacle which has a positively HUGE 32bit address space...  Who could EVER have imagined we'd exceed THAT?
You may as well also include the good old bar codes (such as UPC / EAN13).  What single manufacturer would ever exceed 99999 unique products? Is 99999 unique manufacturers sufficient to cover the entire GLOBE???

Correct me if I am wrong, but my understanding that both the VID and PID are constrained to 16 bits each.  (Thus a range of 0-65535 each at best case)
It's only a matter of time...
(Just like it's only a matter of time before the signed 32bit 'time' counter in original *nix systems 'rolls over'...  Jan 19 2038 IIRC Sometime around 03:19)

Unlike IPv4, USB VID usage isn't even over the halfway point. According to ( http://www.linux-usb.org/usb.ids )

A short filter on that list and we have a total of just 2885 VID's being used and very few are even close to using up their PID space. That represents just 4.5% usage for the entire history of USB and since it costs so much to get a VID people can't just get them that easily. If anyone has access to another more detailed dataset we could even plot the usage and allocation dates roughly but I don't think we are going to be running out of VIDs for quite some time.

IP addresses ideally should be assigned to every device connected to a network so that it is unique across the globe. The people who thought a 32bit address made sense did not realize the potential for everyone to have 2+ devices each with their own IP address. But for USB VID/PID you do not ever going to have a situation that every person on the planet is going to make two or more companies and devices requiring their own VID/PID and $4000+.

[miguelvp]

mute point, FTDI offers 8 PIDs using their VID for free, still allows you to deploy your own driver.
but maybe not for hobbyists

[a210210200]

True, FTDI does offer PID values for your own drivers. I was talking earlier about how functional FTDI clones and larger companies could easily have their own VID without issues.

One example is that most free PID use is contingent on the use of the mfgs product but if you use your own VID you can switch mfg or do whatever basically and there is no need currently for companies to "share" a VID.

[rob77]

A short filter on that list and we have a total of just 2885 VID's being used and very few are even close to using up their PID space. That represents just 4.5% usage for the entire history of USB and since it costs so much to get a VID people can't just get them that easily.

if USB VID is expensive, then what would you call the PCI-SIG fees to get a PCI vendor ID ?

[a210210200]

Relative speaking of course vs. IPv4 address allocation which at first was like handing out candy land jumbo blocks to whoever wanted them until it was too late and they realized that people where sitting on massive unused blocks. (Some nice companies released their unused allocations they were never going to use I think).

[classicalQbit]

This is the first time I post in this forum though i have been lurking for some time.

The company I work for use expensive equipment that is produced in low quantities. The probability that any such equipment manufacturer is able to have absolute control over the entire supply chain at all times for each individual chip is obviously limited.

We have a "air gap firewall" that protect the equipment against a lot of malware, but if a "updated" windows computer transmit an attack, it poses a risk to any equipment it is connected to. Our equipment is interfaced by USB externally, and there is probably a USB to serial interface internally.

To mitigate some of the risks such as FTDI gate:

1) No microsoft equipment or software may connect to any manufacturing equipment. This will probably be a permanent ban.

2) FTDI chips in equipment will be considered a risk. It will be among the factors that will be considered when a supplier wins or looses a tender.

3) All employees have been informed of the risk Microsoft and FTDI forced on us. And that their private equipment also may be at risk.


Hopefully it will be enough. It have been so far. My best wishes for you all and your equipment, may you stay safe.

[8086]

I just got a product change notification email from Mouser, about the FT232, linked PDF: http://www.mouser.com/PCN/FTDI_122757PM11062014.pdf

It's very interesting.

[Macbeth]

I just got a product change notification email from Mouser, about the FT232, linked PDF: http://www.mouser.com/PCN/FTDI_122757PM11062014.pdf

My oh my, Fred Dart please stop digging that hole further! Climb out of it and own up to a moment of insanity.

[SydB]

I just got a product change notification email from Mouser, about the FT232, linked PDF: http://www.mouser.com/PCN/FTDI_122757PM11062014.pdf

Quote from that PCN:
'Although in certain parts of the media it has been implied that there was some form of counterfeit detection algorithm
in FTDI’s latest driver, this is in fact absolutely untrue. There was no mechanism of that description in place and hence
no flagging up of a counterfeit device ever occurred. Exactly the same commands and sequence are sent to a genuine
chip as to a counterfeit chip. Some counterfeit devices simply failed to handle certain commands correctly (again
something that’s shows their lack of suitability for use in serious electronic system design) and they simply end up
quarantining themselves out of harm’s way.'

Does this guy ever ask anyone else for an opinion? What a ridiculous paragraph. 'quarantining themselves out of harm's way'! How can something quarantine itself? I did sympathise with FTDI a little on this issue but releasing a message like this is utterly insulting.

If the sequence of commands in the driver that set PID to 0000 had a legitimate reason other than to affect counterfeit devices (so that the user then detects that they no longer work) then why have FTDI not told us what it is? If that is not a counterfeit detection mechanism then what is? The oddest thing is that after Fred strongly denies there being a counterfeit detection mechanism, the PCN then goes on to say in the very next paragraph that the re-released driver version will contain a detection algorithm!! This man is a lunatic.

Keep digging FTDI. Better still, shut up before you alienate all the engineers that might design in your chips.

[dannyf]

(again something that’s shows their lack of suitability for use in serious electronic system design)

Someone needs to go back to grammar schools,

Although in certain parts of the media it has been implied that there was some form of counterfeit detection algorithm
in FTDI’s latest driver, this is in fact absolutely untrue. There was no mechanism of that description in place and hence
no flagging up of a counterfeit device ever occurred. Exactly the same commands and sequence are sent to a genuine
chip as to a counterfeit chip.

Those two statements are not mutually exclusive.

One can lie by telling 100% truth.

[miguelvp]

Not just that, later on:

This will still uphold our stance against devices that are not genuine.

FTDI has shown itself to be very proactive in combatting the issue of counterfeiting and will continue to be vigilant.
The hope is that the seriousness of the problem is now beginning to be realized and everyone in the electronics
engineering community will do their best to stamp out counterfeiting, as this will benefit the ongoing progression of
technology and encourage future technological advances.

which one is it? no detection? or combat(t)ing and taking a stance against them?

[dannyf]

Looks like the next generation of drivers will indeed have mechanisms to detect fake chips.

[miguelvp]

Yeah, but saying "still" implies the previous driver also did

[electr_peter]

I just got a product change notification email from Mouser, about the FT232, linked PDF: http://www.mouser.com/PCN/FTDI_122757PM11062014.pdf

More lies and (ahem, PR policy) from FTDI officials. They indirectly admit that fakes/clones of all kinds/incompliant devices were damaged on purpose and killer driver was forced silently via automatic windows updates (ups, lawers said we can get sued so we removed it) on equipment not belonging to FTDI. Also, nobody was damaged by this, of course (at least nobody, that had original FTDI chips, that is ).
Some distorted world view. Good luck for anyone trusting company and their products with such management

FTDI (and some forum members) continue to claim that any device that fails with purposely designed killer driver is 100% illegal and breaching their rights, etc.  This is very debatable and almost certainly untrue (FTDI IC implementation of the memory management is purposely/accidentally with bugs. There may be fakes(copies) and clones).
Also, FTDI allowed them to assume that they can damage final user equipment if they want to. This is clearly illegal (and no amount of can justify it).

FTDI, if you are reading this forum, you should understand that final user cannot identify original/fake/clone/copy IC short of de capping IC and checking it. This policy of punishing end user for wholesale issues is short sighted and damaging to FTDI.