loading

Lenovo laptop died!? Any ideas?

I've got a "Lenovo Flex 2-15 Laptop" that recently just randomly shut down and died as if the battery and power was disconnected. I think this might have been caused by a faulty USB device drawing too much current. (specifically, a arduino-based gimbaled motor controller for camera stabilization on my quadcopter) When it was plugged in, the second I clicked to 'connect' to it via a GUI software interface for it, my computer instantly shut down. Since then it shows no signs of life. Any ideas what could have happened and why it does not boot?



It appears to still be under warranty so I'll see what I can do. I really hope that bad things didn't happen. My gimbal still works, but I guess a glitch in the software on the gimbal caused a transient voltage spike or overload on the USB port and either tripped some sort of safety protection or even worse, the mobo failed. Maybe the 5V regulator in the laptop and the one in the gimbal didn't agree with eachother and got into a bit of a voltage war!?!?!? IDK. I have a sneaking suspicion that it is a faulty polyfuse between the battery and MoBo that's failed.



The gimbal has a simple 5V 78m05 SMD regulator between the 12V lipo and 5V logic stuff. However the USB and 5V regulator share the same common connection. I think if something really bad happened, like the voltage regulator failing and dumping 12V into the USB port even for a short time, then I would expect the magic smoke to have escaped from the gimbal and the computer. From my experience, common AVR microcontrollers tend to fail pretty spectacularly with 12V. Since the gimbal still works, I am pretty confident that a low impedance 12V transient had not occurred. I think maybe the gimbal took too big a gulp of current and set off some sort of USB overcurrent protection. But I cannot open the thing, damn modern laptops don't even have removable batteries anymore GRRR! I'll file a warranty claim and see what happens, but does anyone have a better idea what could have gone wrong?

sort by: active | newest | oldest

There is no real chance to damage a USB port if the gimbal thing is working properly and designed the right way.
I had some issues with bad contacts on my 3D printer causing it to try to get the power it needs from the USB port.
This caused the USB port to be shut down right away - had to unplug to get it back.
The failsafe in modern laptops is similar.
You should have multiple ports, at least one is only rated for a max of 500mA - to connect mouse, readers and so on.
On top of this one there should be one or two ports that allow charging on powered devices, like hard drives, robotics and so on.
These ports either supply up to 1amp or up to 2amp if USB3, the new faster USB3 standard allows for up to 3amp.
If anything connected draws more power than specified the port, or the entire laptop will shut down but that should be the only thing happening.

In case your gimbal is faulty somewhere then it would need to send back voltage over the data lines or more than 5V (or reversed polarity) back onto the power lines of the USB connection.
You should check on the gibal if it provides 5V (or more) on the USB port.
There should be nothing when powered up.
Check if the 5V on the gimbal is in any way protected against the USB 5V.
If all is good the best suggestion I have is to add a diode or two to make sure nothing can come from the gimbal onto the 5V and ground connection of the USB port.

-max- (author)  Downunder35m1 year ago
That's about what I figured out. I was able to power a raspberry pi from the port that is supposed to plug into the computer! There shouldn't be any voltage on that connector, and as I guessed, the 5V power rail is shared with the miniUSB connector and the battery. Thing is that it needs 12V for the gimbal motors to function, and 5V for the logic. The software can change the settings of the gimbal in real-time, which is very nice.

So I think the most likely thing is that the power was back feeding into the laptop. I measured the voltage to be 5.016V, so it does not seem like it is high enough to cause problems. This seems more like an intermittent problem though. I did notice that the shell of the miniUSB connector is electrically floating, and I got some strange phantom voltages on it. An ohmmeter test shows it does not connect to anything. :/ Looking at the layout, I am not a fan of the design, the 12V lipo battery is right next to that USB port and the 5V logic stuff with no real amount of creepage distance. The regulator is a ams1117.

I think the best solution might be some sort of optical data link between the computer and the board to make this type of problem impossible.

Yes, there are several opto-isolating USB solutions out there. Analog Devices make a chipset - its not a trivial task opto-USB, but I found it less expensive to buy an Olimex pre-built board - it cost the same as the one off chips.

-max- (author)  steveastrouk1 year ago

I figured it might be difficult, I don't know much about differential pairs like the one used for USB, the logic level or anything. Are the differential signals voltage referenced to the power rails? Like they would be if driven via a H bridge (a push-pull output for each part of the differential pair)?

I would probably not do it on the USB side, but on the UART side because the microcontroller is UART based and has a FTDI-like part to interface the microcontroller to the computer. I have tons of optoisolator chips that would probably work fine for this.

I think it is the motors that cause the problem, if they are not properly isolated from the USB side all things can happen.
But why not try a powered USB hub first?
Maybe not the 2$ from Ebay though...
The one I use got a little modification:
I disconnected the 5V rail coming from the computer USB entirely on the hub.
A switch was added to bridge this if required.
So all the data handshake stuff and so on can only happen if power is connected to the hub or if the switch is activated.
Of course I don't know if your voltage problem might come over the data lines...
In that case you would have to check the hub for proper protective features.

As a last resort to prevent voltage spikes on the 5V side of the USB connection you could add a beefy 5.3 or 5.5V zener diode.
But again this won't help for problems on the data lines.
Maybe a nice little filter would help too but you would need to check first what type of "dirt" causes the problems on the 5V rail.

-max- (author)  Downunder35m1 year ago

Well it seems after I let the battery fully discharge for almost a week, something inside (like a soft-latching overload detection circuit) has finally reset and the computer booted up after applying power! :D I suspect that I just pulled a little too much juice from the 5V rail, or maybe the regulator on my gimbal was back feeding power which tripped that overload detection circuit, I don't know. But from now on I am making a isolated USB interface for it.

Yes, you can do it on the UART side, but an isolated USB interface is a useful thing for the armoury.

I don't know how to fix your computer, but can say I had a similiar problem when using a ESP8266 microcontroller. It draws too much current and must be powered externally from your PCs laptop. Of course I tried anyway and experienced the exact same thing you are describing, but luckily in my case the PC came back to life. It was due to more current being drawn than the port can handle. That could have been your problem as well, but perhaps yours was even worse and it fried the machine.

-max- (author)  siliconghost1 year ago

Well it seems after I let the battery fully discharge for almost a week, something inside (like a soft-latching overload detection circuit) has finally reset and the computer booted up after applying power! :/ :D I suspect that I just pulled a little too much juice from the 5V rail, or maybe the regulator on my gimbal was backfeeding power which tripped that overload detection circuit, I don't know. But from now on I am making a isolated USB interface for it.