Filtering ESD for CE compliance

Log in or register to post comments
Go To Last Post
7 posts / 0 new
Author
Message
dsandber
dsandber's picture
Level: Hangaround
Joined: Wed. Feb 11, 2004
Posts: 119 View posts
#1
Posted by dsandber: Sun. Jun 13, 2010 - 01:15 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've been trying to filter ESD spikes using caps, ferrite beads, TVS diodes, capacitors, and schottkys but haven't succeeded. Details:

I was told that using a modified BBQ lighter (has a piezoelectric spark generator) was a good way to simulate the ESD testing needed for CE compliance.

I am powering a commodity motherboard from a commodity power supply that is in turned powered by a +32V input. When I generate a spark on the +32V input (or the GND) the motherboard reboots or locks-up.

I CANNOT mechanically prevent the +32V/GND from being exposed to the ESD spark due to the design of the device. I MUST find a way to filter the ESD spark so the power-supply doesn't see it.

Because my device is battery powered, I can't rely on the presence of an earth ground either. So I need to somehow filter a spark (crosses a 1cm air gap in 30% humidity) so that the power-supply (and everything downstream) sees only 32V DC.

Everyone tells me this is a black-art and that even a sufficiently fast scope (1GHz+) won't be much help in designing a solution.

I'm hoping some of you "wizards" can give me a starting place to address this "black art" since right now I'm just trying random components I have lying around with little luck :)

TVS, Varistor, Sideacter, MOV, MLV, caps, and in what combo? I know its trial and error, but having a good place to start could speed things up immensely!

Thanks!

-Dan

Tags:

Learning and Information, General Electronics
Top
leon_heller
leon_heller's picture
Level: Raving Lunatic
Joined: Fri. Jul 27, 2001
Posts: 7680 View posts
Location: St. Leonards-on-Sea (UK)
Posted by leon_heller: Sun. Jun 13, 2010 - 01:51 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I used AVX Transguards on a system I designed where I used to work. It passed the CE tests without any problems.

Leon Heller G1HSM

Top
ka7ehk
ka7ehk's picture
Level: 10k+ Postman
Joined: Fri. Nov 22, 2002
Posts: 20511 View posts
Location: Tangent, OR, USA
Posted by ka7ehk: Mon. Jun 14, 2010 - 03:33 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You have to "filter" ahead of the supply, on the 36V input. And you MAY not be able to fix it if you are stuck with a commodity power supply. For example: If the ESD drives the input to ground, or below. the supply MAY shut off. In this case, down-stream from the transient protector, you might need a series diode and a big cap that can supply the power supply during the negative ESD transients.

But, you can only tell if this is happening by observation, inference, and more testing.

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

Top
oddbudman
oddbudman's picture
Level: Hangaround
Joined: Thu. Jan 16, 2003
Posts: 342 View posts
Posted by oddbudman: Mon. Jun 14, 2010 - 04:54 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I agree with Jim.

I think you should check out the capacitance you have. TVS diodes can help to clamp levels to safe limits. Layout will be another thing to check.

You need to protect the board at the inputs, putting the protection components more central to the pcb will simply expose more of your tracks to the ESD event.

Are you sure the power supply is the cause of the reset? Just becuase you are applying the ESD event there does not mean that it is necessaraly the problem area. The ESD event may be getting to other areas (like the reset line) of the board.

oddbudman

Top
Kartman
Kartman's picture
Level: 10k+ Postman
Joined: Thu. Dec 30, 2004
Posts: 22788 View posts
Location: Melbourne,Australia
Posted by Kartman: Mon. Jun 14, 2010 - 12:13 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Regardless of CE comliance, it seems your system is rather sensitive and would give problems in the field. I agree with oddbudman - the ESD event is travelling down some wires and finding its way back through a path that doesn't appreciate it. Cut that path off before it gets too far!

Is the enclosure conductive? Metal? Is the 0V connected to it? For higher current power stuff I tend to use varistors as they can usually take more of a kick - they also explode if you kick the too hard so be sure to have a fuse or suchlike to protect it. Make sure all other devices connected to the boards have the 0V signals connected to the enclosure - this ensures that the cirrent should pass through the enclosure, no the boards.

Top
dsandber
dsandber's picture
Level: Hangaround
Joined: Wed. Feb 11, 2004
Posts: 119 View posts
Posted by dsandber: Mon. Jun 14, 2010 - 07:15 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks everyone who wrote. I have some comments, questions, but first, some progress!

We noticed that the same reset/lockup problem happens when we put the shock into a wire that is nearby but not electrically connected to the power-supply. So apparently our problem is more radiated than conducted.

I stuck 5 ferrite beads of various sizes on the power supply input cable, and the ESD immunity improved massively.

I experimented and it seems that the hugest ferrite bead closest to the spark source, with the power cable wrapped around it four times (three does not work as well) also does an excellent job.

Seems that I don't need the TVS and regular diodes (for reverse protection) because the power-supply has a TVS built-in which apparently is sufficient.

So I'm quite excited to at least have a solid lead here.

Given this new knowledge of the problem, is there anything I can do that is more elegant (and lighter/cheaper) than a huge ferrite ring?

Jim, I'm worried about the PSU "starving" too, but there is a battery in parallel (with a diode so the PS doesn't feed into it) so hopefully this will prevent that.

Any thoughts on TVS vs. MOV?

Yes, the entire PSU and motherboard is enclosed in a metal box, and we are working on fine tuning the box for CE immunity/emissions. This will obviously also be a huge help for the ESD.

The motherboard we are using is mounted with PEMS to the aluminum box. Assuming the PEMS have been coated with Alodine, is this a sufficient connection between system ground and chassis? Should we have multiple connections or just one?

Top
Geoff
Geoff's picture
Level: Hangaround
Joined: Mon. Apr 24, 2006
Posts: 285 View posts
Location: Sydney, Australia
Posted by Geoff: Wed. Jun 16, 2010 - 05:52 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

dsandber wrote:
I was told that using a modified BBQ lighter (has a piezoelectric spark generator) was a good way to simulate the ESD testing needed for CE compliance.

Yep, used this before myself for the same purpose. I even had a council cleanup in my neihgbourhood at the time so I canvassed a few dumped BBQs and got the ignitor for free. Duplicated the problem they had found in CE testing, found an easy fix and passed the formal tests. Isn't it great when a low tech method is useful :D

Top