USBLC6-2SC6 ESD wont protect

[Laszlo Nagy Kanasz]

Posted on August 24, 2017 at 09:47

Hi

We have a device which is an USB powered(USB2.0), and on the same USB port is the communication with the PC.

So it is powered from the PC.

For the ESD protection of the USB lines we add the USBLC6-2SC.

Our processor is a TIVA TM4C1294NCPDT type.

The device is continually streaming, the PC program only captures what the device is sending.

The problem is with the ESD events.

We tray in laboratory, and a 2kV ESD spike is enough to make problems.

As we see the device wont stops, nor even reset, so the ESD spike is not interrupting the streaming, but the PC detect a disconnect of the USB device, and after it connects back automatically...(but we have to restart our capturing program

every time)

Is this a specific problem related to our HW design, or the problem is with the ESD protection chip itself.. I read on some posts that this ESD TVS array is not the best for the cases when the VBUS status is detected... but our case is not that.. we does not pull off/on the USB plug, and the power line (VBUS) is working continually (we use some galvanic separation DC/DC to power our processor part from the VBUs line).

Is there some issues using this ESD protection ?

Should we use some other type, other methods?

Part of our USB power/communication schematics section is attached.

Best Regards

Laszlo

[AvaTar]

Posted on August 24, 2017 at 10:34

Our processor is a TIVA TM4C1294NCPDT type.

Not sure if the ST forum is the right place, then ...

As we see the device wont stops, nor even reset, so the ESD spike is not interrupting the streaming, but the PC detect a disconnect of the USB device, and after it connects back automatically...(but we have to restart our capturing program

every time)

As far as I know, this is acceptable behaviour in a burst/surge test.

The device shall just not break.

However, you could improve your firmware to detect transmission errors and/or disconnects, and act accordingly.

But that would probably be a question to the Tiva TM4C forum.

[waclawek.jan]

Posted on August 24, 2017 at 11:18

ST forum is OK, but

https://community.st.com/community/emi-filtering-and-signal-conditioning

.

As ESD is essentially high-frequency RF, much depends on the layout. The DS deals with it.

If you don't connect the VBUS terminal directly to the connector's VBUS, you render more then half of the device ineffective.

JW

[Laszlo Nagy Kanasz]

Posted on August 24, 2017 at 10:41

Hi

The question  was more about the used ESD safety part, the USBLC, than then the used processor type.

As the USBLC is an ST component, that's why i think to ask questions about this part on this forum.

[Amel NASRI]

Posted on August 24, 2017 at 12:01

For sure this thread shouldn't be in the STM32 forum.

We check where it has to be moved.

[AvaTar]

Posted on August 24, 2017 at 12:14

As ESD is essentially high-frequency RF, ...

Not sure if we talk about the same. I meant

E

lectro

S

tatic

D

ischarge.

But there is always lots of space to interpret those 3 letter acronyms ...

[waclawek.jan]

Posted on August 24, 2017 at 12:53

Not sure if we talk about the same

Yes we do. While most people think of ESD as 'current-limited high voltage', its proper treatment has to take into account the accompanying very fast pulse edges, converting into an extremely wide frequency spectrum.

JW

[AvaTar]

Posted on August 24, 2017 at 13:24

True, the pulse-like discharge has a wide spectrum, usually far into the MHz range.

While such a discharge would also affect nearby elements via radiation, the galvanically connected elements usually take the hardest 'hit', i.e. most energy.

And protective circuitry on a GPIO might blow, without any visible indication.

BTW, I once blew a 3.3V chip with mains voltage. THAT left clearly visible indications ...

[Eugene Solo]

Posted on August 24, 2017 at 16:39

Those devices are meant to protect from hardware failures, not from software glitches. ESD spike will be dumped on internal TVS structure and will not burn your processor instead. While ESD event is on, ESD protection device 'looks like' short circuit on bus. So signal integrity will be lost for that short period of time and how your software will recover from this - nobody knows (except you, ofc )

[Laszlo Nagy Kanasz]

Posted on August 24, 2017 at 14:41

Hi

As I read on some other posts, there are various type of connections regarding the ESD TVS safety elements.

Some of them connect the chips GND onto the USB connector shield. ST mention connecting the USBLC6-2SC onto the GND line of the USB connector.

I try it both, and also try this combinations with the VBUS leg lifted up on the USBLC.

But the situation is the same in any combinations.

So basically what I can see is that the device continually streaming, but the PC hangs OFF the communication, and redetect again the device after the ESD event.

So my question here is: Is it a normal situation with the USB? Could any of the safety solutions give a safety level in which does not comes to this communication error?

Just to know am I looking for some solution which does not exists, an the whole story is only about to prevent burning out the device elements, and not to enable a continuous communication.

So basically the communication error is a normal situation, and whatever we do it will present?  and the issue is only how to prevent the instrument/pc electronic from burning out...

Best Regards