ESD and STM32F407

[VE7HR]

Posted on March 05, 2014 at 01:44

It's ESD season once again. I have been getting far too many STM32-F407 TQPF IC's being blown up. I believe ESD is the culprit. Failure mode is a bit used as an input is blowing out the internal pullup or that what it looks like. The pin on a bad board is sitting a 0V where a good board is pulled up.

The pins are being used as a rotary encoder input. Pin PC5,PC6,PC7 and PB4, PB5, PB6. From what I can read the part is a class 2 device for ESD. What is disconcerting is I am still getting failures after I tack PESD0603-24 Diodes across the pins to  ground at the encoder. This should solve the issues but it does not seem to. They are specifically designed as ESD diodes. I have added on the next run of PCB's NUF4402MN devices but have many boards out there to support. Any suggestions are appreciated. I see, to be missing something. Thanks Dave #esd

[francescatodiego]

Posted on March 05, 2014 at 08:14

Hi Dave

The

board has passed the

EMC test

?

Happening in the

field

or

are you still doing

tests

? The

electrostatic discharge

test can

be

easily proved

.

Are you sure

that

is

not a

problem

of

another

kind?

With my

(little)

experience

I don't see

much to be done

if the input

can not stand

the spike

you must

isolate

for example

with

something like the

26ls33.

This also help if you connect encoder with long cables (I have used for encoder interface installed in rolling mill plants)

[jpeacock2399]

Posted on March 05, 2014 at 15:58

I see a couple of problems.  First, you are using 24V devices for pins rated only to 5V.  Second, where's the ESD protection for the power to the encoder?  According to TE the PESD0603 devices aren't rated for protecting power rails.  And where are the devices being placed?  They should be inline between the connector and your inputs.

I use the ST USBLC6-4 diode arrays on my encoder connectors.  The array clamps the three inputs (A, B, Index) to the 0V and 5V rails, plus the USBLC6 protects the 5V rail to the encoder with a fast breakdown zener.  You might look at USB oriented ESD protection devices.  They are designed for high frequency, 5V environment common to many rotary encoders, and there are both 2 and 3 line versions plus power rail protection.   Jack Peacock

[VE7HR]

Posted on March 05, 2014 at 17:11

Jack,

Some good points. 1) The rotary encoders I am using are Bournes PEC11 devices which are mechanical switches so there is no external power involved. Just switches when closed pull down the inputs in quadrature. 2) The Encoders are mounted direct to the PCB. 3) In the test boards the ESD devices are mounted right on the pins that the encoders are mounted to the board. So they are mounted at the input. The pins are 3 in a row with the middle on being signal ground which is directly to the power plane. Can't get much more direct. 4) Yes they are 24 V device but looking at that series of device even though they are different voltage rating clamp at the same voltage. A closer examination seem to indicate they trigger at a level below the class 2 specified voltage and all clamp at the same voltage. They don't specify the peak voltage just trigger and clamp 30ns later. 5) I think I will see if I can find another device that triggers at a lower voltage. I have added  NUF4402MN in the next board that look really good but are hard to retrofit. They are a dual ESD diodes with series resistor in the Pi config. They should be good. 6) I do use the USBLC6 on my USB and have had no issues. That is another thing to look at. For the retrofit I need to find a 0603 package single device to solder between the pins that is available and limits the peak ESD event to the class 2 limits which depends on how you read the spec 500-2000V. Any suggestions in the short term?? Thanks Dave

[VE7HR]

Posted on March 05, 2014 at 17:17

The project is in field testing. The preliminary EMC and ESD tests seemed acceptable. We have a small number in lab and in field. Getting common failures from a couple of software developers or in uncontrolled real environments. I am rather sure its ESD after the post failure interviews. And its limited to a single input pin in fact we have two failures and the failures are limited to just 2 of the 6 possible pins.

Thanks for your good ideas. Dave

[VE7HR]

Posted on March 05, 2014 at 17:27

Jack,

After doing search anotherway at my favourite supplier I think I might try out the following devices to see if they work better. They seem to have better documentation

VCUT05D1-SD0-G4-08. the other one might be too slow to clamp they seem to have very sparse documentatation.

Thanks Dave