FTDI drivers totally crash my PC

[Monkeh]

point in fact is : a hairball like that with high speed things like USB and RF stuff like wifi is asking for it.

Maybe, but  that's the way most people will be using those modules.

So it's FTDI's fault that people take a load of cheaply designed and manufactured piles of crap and randomly plug wires into them?

[Fungus]

So it's FTDI's fault that people take a load of cheaply designed and manufactured piles of crap and randomly plug wires into them?

According to FTDI the reason people are cloning their chips is because their drivers are soooo good.

(That was one of their excuses for bricking everybody's chips...)

[dannyf]

So it's FTDI's fault...

Whether it is FTDI's fault or not doesn't matter here.

What matters is to blame FTDI for whatever fault there is.

[Fungus]

What matters is to blame FTDI for whatever fault there is.

Yes, that's the only reason I swapped out the FTDI chip that I was happily using in my project

(and many other projects before that).

[zapta]

So it's FTDI's fault that people take a load of cheaply designed and manufactured piles of crap and randomly plug wires into them?

'Cheaply designed', 'piles of crap' and 'randomly plug wires' are your own subjective labels, not facts. Not a good practice if you expect serious answers.

[zapta]

What matters is to blame FTDI for whatever fault there is.

They brought it on themselves.

[Monkeh]

So it's FTDI's fault that people take a load of cheaply designed and manufactured piles of crap and randomly plug wires into them?

'Cheaply designed', 'piles of crap' and 'randomly plug wires' are your own subjective labels, not facts. Not a good practice if you expect serious answers.

Out of this thread, I long ago stopped expecting anything but pointless, bandwagon FTDI bashing.

[dannyf]

There was zero attempt made to find out what the issue was;

There was every attempt made to pin it on FTDI, deservedly or not.

[Fungus]

There was zero attempt made to find out what the issue was;

I already told you why I'm not doing it this week - I need my PC to keep working, I don't need any weird crap happening that might take out the motherboard, corrupt the disk, or anything else like that. Feel free to donate another PC if you're in such a hurry.

There was every attempt made to pin it on FTDI, deservedly or not.

Even if it's deserved it'll be because my adapter is some cheap piece of garbage made in China, right?

[zapta]

Out of this thread, I long ago stopped expecting anything but pointless, bandwagon FTDI bashing.

Pointless? Even FTDI's CEO realized he was wrong and reverted that driver change decision.

[miguelvp]

Out of this thread, I long ago stopped expecting anything but pointless, bandwagon FTDI bashing.

Pointless? Even FTDI's CEO realized he was wrong and reverted that driver change decision.

So what is the point now?

[zapta]

[qSo what is the point now?

Beware of FTDI, hide your children. Why do you ask?

[wraper]

Even if it's deserved it'll be because my adapter is some cheap piece of garbage made in China, right?

Even if it did not deserve it'll be because it is FTDI, right?

[Fungus]

Even if it did not deserve it'll be because it is FTDI, right?

Ah, I think I see your problem...

Answer:

No. I'd stop using any brand of device that repeatedly locked up my PC. And I'd call them on it if they were claiming that the thing that distinguishes them from their competitors is "quality drivers".

[wraper]

This is not FTDI branded device. This is One hung low adapter with a chip calling itself ftdi. It might be counterfeit chip as well, if new driver don't kill it, doesn't mean that it is genuine. I'm aware of at least 2 types of clones: Chinese supereal and Belarus from Integral factory. But I don't think that it is even fake chip at fault but crappy schematic of that adapter and your super solid ground combined with another woodoo things. You even didn't answer if your pc crashes with driver not installed or not. IMO this is hardware fault that crashes the driver or entire usb hub.

[Fungus]

This is not FTDI branded device. This is One hung low adapter with a chip calling itself ftdi.

Yep, and this is what you'll cling to, no matter what I do....

[janoc]

Hello guys,

I think that whether or not the FTDI adapter is genuine or not is likely a red herring here. It is fairly easy to verify whether or not it works - if it tested out as "Genuine", then no matter whether or not it is really genuine, we can be sure that the machine communicates with it just fine.

I have the same Wifi module and I was testing them just two days ago with a FT2232H breakout board wired up with loose wires on my bench and it works just fine like that - so you can rule out the wifi interference as the likely cause too. Anyhow, that shouldn't cause a PC crash neither.

I have noticed two things:

1) Lack of ground connection between the FTDI adapter and the Arduino - bad bad, that will make things to not work correctly. But that shouldn't cause a crash.

2) Where is the whole shebang getting its power from? Through that single USB connection? If yes, then it is very likely that you are right at the edge of the power budget of the hub this is connected to. Those wifi modules are really power hungry, they require about 300mA of current by themselves. Add the current required by the power regulator, level converters on your gadget, the USB2serial bridge and whatever else you may have connected to the same hub (hint - mouse, keyboard ...) and you are easily running out of power (remember, USB specifies only 500mA per port from a *powered* hub!). Is this a laptop? That would make such power problems much more likely.

When you change the USB2serial modules, you exceed the power budget, the overcurrent protection of the hub kicks in and shuts things down. Now depending on the hub, it might shut down only the offending port, but many will shut the entire hub down, because it is simpler to implement (e.g. via a simple polyfuse). If this is a hub that is used to connect your keyboard and mouse too, then boom - your machine seems dead, Windows doesn't expect such situation neither and can go haywire, etc.

[Fungus]

I have noticed two things:

1) Lack of ground connection between the FTDI adapter and the Arduino - bad bad, that will make things to not work correctly. But that shouldn't cause a crash.

There's a ground via the USB cable - USB is always tied to mains ground (if available).

I tried an extra ground wire, it made no difference.

2) Where is the whole shebang getting its power from? Through that single USB connection? If yes, then it is very likely that you are right at the edge of the power budget of the hub this is connected to. Those wifi modules are really power hungry, they require about 300mA of current by themselves.

Yes, I know about the power requirements.

My USB ports can deliver 800mA each. USB ports and Arduinos have fuses in them against short-circuits/overloads.

Here's the Arduino schematic: http://arduino.cc/en/uploads/Main/arduino-mega2560_R3-sch.pdf

The fuse is the thing marked "F1" right next to the USB connector. It's a Bourns MSMF050-2 (500mA)

If this is a hub that is used to connect your keyboard and mouse too, then boom - your machine seems dead, Windows doesn't expect such situation neither and can go haywire, etc.

They will only 'die' until the fuse resets. There's no reason for Windows to crash, USB is designed for this. It's not just the USB, one time I had an EEVBLOG video playing when I pressed reset and it froze too. My keyboard is an old fashioned Model M that goes through a serial connector, not USB.

(PS: I'm pretty sure I've overloaded USB a couple of times over the last few years but I don't remember locking up the PC)

[janoc]

There's a ground via the USB cable - USB is always tied to mains ground (if available).

I tried an extra ground wire, it made no difference.

That ground via the mains earth is going to be extremely noisy. Furthermore, you are creating a ground loop, that will make the signal integrity even worse. It could even blow something up if there is a sufficient potential difference between the USB connectors (Arduino+your FTDI thingie) - e.g. if you decided to have the connections going to two different devices (such as a PC for the FTDI and a USB charger powering up the Arduino) connected to two different power outlets. The mains earth wiring may not even be connected on one side (e.g. the USB charger) or there could be easily tens of volts difference between the two outlets (especially in older buildings the wiring can be horrid!). In short, this is a major no-no and asking for big problems. Dave has a good video on the perils of mains earth wiring and how not to blow up your instruments, I strongly recommend you to check it out.

However, I agree, that should not cause a Windows crash like that, if things didn't go bang already then at worst you won't be able to talk to the chip.

Yes, I know about the power requirements.

My USB ports can deliver 800mA each. USB ports and Arduinos have fuses in them against short-circuits/overloads.

Here's the Arduino schematic: http://arduino.cc/en/uploads/Main/arduino-mega2560_R3-sch.pdf

The fuse is the thing marked "F1" right next to the USB connector. It's a Bourns MSMF050-2 (500mA)

Are you sure that the USB ports are actually able to deliver that current in your setup? If there are other things connected to the same hub already, they might not be. A good thing to make sure would be to connect this setup through an external, powered USB hub instead of directly to the PC. Alternatively, you can always power up the Arduino from a wall wart and get the 3.3V you need from its power regulator. That should be good enough.

The part about the polyfuse on the Arduino is not relevant - I was talking about the total current consumption at the point where this is connected to the PC - i.e. the USB hub. That includes all other hardware you may have connected to the same hub - e.g. a keyboard or your mouse. The polyfuse not tripping only tells you that the setup (Arduino + Wifi module) likely consumes less than 500mA (+- some tolerance, also polyfuses don't catch short power spikes). It doesn't "see" any current taken by the rest of your hardware that isn't connected through it - your FTDI adaptor, anything else connected to the same hub, etc. If you want to be sure, hook up your multimeter and measure the current through both the Arduino and the FTDI bridge USB connections (you may have to cut a trace on the board, cut up a cable or use a breakout box like this one:
http://archive.siliconchip.com.au/cms/A_112479/article.html).

They will only 'die' until the fuse resets. There's no reason for Windows to crash, USB is designed for this. It's not just the USB, one time I had an EEVBLOG video playing when I pressed reset and it froze too. My keyboard is an old fashioned Model M that goes through a serial connector, not USB.

I have an external hub that won't reset by itself after the short/overload is removed but you have to actually powercycle/disconnect it to get the dead port back - it uses a more advanced power management  solution than a simple polyfuse. Another one built into my monitor will stop working completely after a buggy device (i.e. not overload, it is a project I am working on) is connected and I have to powercycle it before the machine recognizes it again. So don't rely on it recovering by itself.

Re Windows crashing - if the hub is a motherboard hub, then Windows may simply not expect it to go offline - and then all bets are off. I have seen BSODs in such situations ... Could you check in the event log whether there is anything there after you restart the machine? Normally there should be an error logged if there was a problem. Also, it would really help if you could show the entire setup how is it connected to your PC, like this we are left guessing.

(PS: I'm pretty sure I've overloaded USB a couple of times over the last few years but I don't remember locking up the PC)

Yeah, it is weird. Try the setup with a powered external USB hub if you aren't using that already or a completely external power supply, without relying on USB power at all. That will allow you to rule out the power issues.

I have a powered USB hub on my bench for all USB connected hacking for that reason, I don't want to bust my mainboard or short the PC's PSU if I do something stupid like shorting the 5V rail by accident or sending 12V on the data lines by mistake.

[sunnyhighway]

Mouse cursor freezes, caps lock key doesn't make the keyboard light change, video stops playing.

They will only 'die' until the fuse resets. There's no reason for Windows to crash, USB is designed for this. It's not just the USB, one time I had an EEVBLOG video playing when I pressed reset and it froze too. My keyboard is an old fashioned Model M that goes through a serial connector, not USB.

These two data-points are solid proof the FTDI drivers are at fault.

Microsoft clearly states that a kernel-mode driver MUST return control to the kernel within 100 ms when it get a time-slice. Regardless if it needs more time or not in order to do what that specific kernel-mode driver needs to do. This rule must be obeyed in order to guarantee responsiveness of the operating system.
USB has no h/w interrupt, it can only do polling. Therefore it is impossible that interrupts are causing the OS to freeze. This makes the USB driver in fully responsible for locking up, regardless wheter the FTDI chip is talking rubbish simply refuses to answer, or completely swamps the USB port. I don’t even care what caused it, whether that being a floating ground, rfi or alien x-rays. This is the reason why kernel-mode drivers are littered with code for time-outs, just to obey that simple 100 ms rule. Forget one and you have a potential cause for a system wide freeze when the driver is waiting for something that may never happen.

Microsoft also states that kernel mode drivers MUST take extreme care not to allow for any buffer overrun, as kernel mode drivers have access to the full address-space and CPU registers. One tiny mistake and it could easily cause an endless loop or BSOD.

[marshallh]

During 2011 I discovered a case where under certain conditions unplugging a FT2232H based device with an outstanding endpoint transaction caused a BSOD on Win32. The problem has long been solved but weird cases aren't impossible. In OP's case I would suggest further testing on multiple computers to rule out the issue.

[Fungus]

That ground via the mains earth is going to be extremely noisy.

It's not via the fusebox in the basement, it's two USB cables plugged into the same PC. The maximum distance is a couple of meters. I wouldn't do it for (eg.) high-speed SPI but it works fine for RS232. The fact that I can buzz continuity from my Arduino to any mains socket in the house is just a bonus.

The part about the polyfuse on the Arduino is not relevant

If the USB ports can deliver 800mA and the Arduino's fuse is 500mA then I'd say it is. The Arduino should shut down first...

Could you check in the event log whether there is anything there after you restart the machine?

"The previous system shutdown at 12:57:50 PM on ?11/?3/?2014 was unexpected."
"The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly"

Nothing else.

[Fungus]

Mouse cursor freezes, caps lock key doesn't make the keyboard light change, video stops playing.

My keyboard is an old fashioned Model M that goes through a serial connector, not USB.

These two data-points are solid proof the FTDI drivers are at fault.

Good catch! I hadn't connected those two dots.

[HackedFridgeMagnet]

Mouse cursor freezes, caps lock key doesn't make the keyboard light change, video stops playing.

My keyboard is an old fashioned Model M that goes through a serial connector, not USB.

These two data-points are solid proof the FTDI drivers are at fault.

Proof might need a bit more than those two data-points.
Those two statements dont even mention the FTDI drivers being loaded.


Maybe Fungus could start with the driver version and the OS version.
Then the FTDI board manufacturer.
Then he could say how he is so sure the IC is from FTDI.

It's a bit strange to be 5 pages into this thread and still basic but relevant information like this hasn't been given.

[Fungus]

Then he could say how he is so sure the IC is from FTDI.

It's a bit strange to be 5 pages into this thread and still basic but relevant information like this hasn't been given.

Precisely because there's no way to be 100% sure of this. As the tread has progressed it's become pretty cleat that no matter what I do/say, the naysayers will say it came from eBay therefore it must be fake/crap, therefore I'm at fault. QED.

And if it's not that it will be because my motherboard is causing the problem, not FTDI's driver, etc., etc. None of which can be proven either way.

...all the time ignoring that fact that swapping FTDI out of the equation without touching anything else made a consistent/repeatable problem vanish.

Rant over. Here's some info. Be sure to let us know what insightful information you got from it: