USB data line protection - is it needed?

[László Monda]

Hey guys,

I wonder what you think about the necessity of USB6B1 part -
http://www.st.com/stonline/books/pdf/docs/5605.pdf

I assume that laptops / PCs should have some similar parts inside of
them and putting it into a device is overkill. Anyways, I don't have
a clue about analog issues and I'm just guessing so I'm interested
about your opinions.

Thanks.

--
László Monda <http://monda.hu>

[Bob Paddock]

On Fri, Apr 8, 2011 at 3:58 PM, László Monda <la...@monda.hu> wrote:

> I wonder what you think about the necessity of USB6B1 part -
> http://www.st.com/stonline/books/pdf/docs/5605.pdf
>
> I assume that laptops / PCs should have some similar parts inside of
> them

They should. Some manufactures may skimp to cut costs. Hard to prove.

> and putting it into a device is overkill.

Not at all. The device needs protection as well.

> Anyways, I don't have
> a clue about analog issues and I'm just guessing so I'm interested
> about your opinions.

Static protection should always be part of the USB hardware design.
I gave an introduction to static issues in my blog:

http://blog.softwaresafety.net/2010/04/software-bug-or-electrostatic-discharge.html

[Opendous Support]

You should add some sort of ESD protection, but not necessarily
something that expensive. A lot of charge can build up on a PCB and
it can all flood into your IC's delicate analog USB PHY if you don't
add ESD protection. However, if your board will require a USB hub IC
or USB switch you should check the corresponding datasheet as most
already have significant ESD protection.

You can use simpler and cheaper bidirectional ESD suppressors. They
simplify routing as you only need to connect them between the data
lines and the connector shield (the bulk of your USB connector). 0603
size are DIY'able and widely available. The key specifications are a
5V to 12V operating voltage and low capacitance.

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=CG0603MLC-05LECT-ND
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=ESD03A5V5R17VCT-ND

I have posted this before but you should also look over the following
for notes on USB peripheral design. Pg.9 and 12 specifically.
http://www.ti.com/sc/docs/apps/msp/intrface/usb/emitest.pdf

> --
> You received this message because you are subscribed to the Google Groups "LUFA Library Support List" group.
> To post to this group, send email to lufa-s...@googlegroups.com.
> To unsubscribe from this group, send email to lufa-support...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/lufa-support?hl=en.
>
>

[László Monda]

Thanks for providing all the useful info, guys.

What I understand based on the earlier information is that there are
various solutions of various costs and efficiency. My question is
that how much overcurrent / ESD is a problem in the real life and how
much should I worry about it. Let's say it affects only 1 people out
of 10,000 in general in which case I wouldn't worry too much.

Also, I guess there are various applications for which it is critical
for some kind of protection to be implemented, for example space,
military or medical applications. What about less stricter product
categories like consumer products or developer boards? Does the
Arduino have any kind of protection for example?

--
László Monda <http://monda.hu>

[Steve Riedl]

I used an STF202-22 on my board. It has overvoltage, ESD, RF AND the two
22 ohm resistors in a package about the size of the 2 22ohm resistors
together. I think they were switching to Rohs and I had a hard time
getting the parts a year or so ago, but they should be readily available
now and are cheap as well.

--
Steven Riedl
rie...@gmail.com
(720) 279-4895 (Google Voice)

"The difference between genius and stupidity is that genius has its limits" - Einstein

[Opendous]

>What about less stricter product categories like
>consumer products or developer boards?

You are most likely to encounter IC-damaging situations with
consumers. ESD isn't a consumer safety issue, it is a device safety
issue. In general you are correct that it is probably a 1 in 10000
connections issue but heat and dry conditions increase the likelihood
of ESD events (as do carpeted environments, wearing wool, or other
static generating situations). The Arduino UNO uses PGB1010603MR ESD
Suppressors:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=F2594CT-ND

>STF202-22

I was unaware of such an all-in-one IC and it is indeed a perfect
solution:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=STF202-22T1GOSCT-ND

However, if you are not sure if you need the extra cost, use 0603
ESD suppressors and then simply don't populate them if you decide not
to use them. They are not in-line/series so they won't affect
routing.

In single quantities all the noted solutions cost about the same but
at QTY-1000:
CG0603MLC-05LE 0.093 * 2
ESD03A5V5R17V 0.096 * 2
PGB1010603MR 0.257 * 2
STF202-22 0.365 * 1

The STF202-22 is much better value compared to the two ESD
suppressors on the Arduino.

On Apr 8, 11:21 pm, Steve Riedl <ried...@gmail.com> wrote:
> I used an STF202-22 on my board. It has overvoltage, ESD, RF AND the two
> 22 ohm resistors in a package about the size of the 2 22ohm resistors
> together. I think they were switching to Rohs and I had a hard time
> getting the parts a year or so ago, but they should be readily available
> now and are cheap as well.
>
> On 4/8/2011 3:16 PM, L szl Monda wrote:
>
>
>
> > Thanks for providing all the useful info, guys.
>
> > What I understand based on the earlier information is that there are
> > various solutions of various costs and efficiency.  My question is
> > that how much overcurrent / ESD is a problem in the real life and how
> > much should I worry about it.  Let's say it affects only 1 people out
> > of 10,000 in general in which case I wouldn't worry too much.
>
> > Also, I guess there are various applications for which it is critical
> > for some kind of protection to be implemented, for example space,
> > military or medical applications.  What about less stricter product
> > categories like consumer products or developer boards?  Does the
> > Arduino have any kind of protection for example?
>
> > On Fri, Apr 8, 2011 at 10:47 PM, Opendous Support

> > <opendous.supp...@gmail.com>  wrote:

> >> You should add some sort of ESD protection, but not necessarily
> >> something that expensive.  A lot of charge can build up on a PCB and
> >> it can all flood into your IC's delicate analog USB PHY if you don't
> >> add ESD protection.  However, if your board will require a USB hub IC
> >> or USB switch you should check the corresponding datasheet as most
> >> already have significant ESD protection.
>
> >> You can use simpler and cheaper bidirectional ESD suppressors.  They
> >> simplify routing as you only need to connect them between the data
> >> lines and the connector shield (the bulk of your USB connector).  0603
> >> size are DIY'able and widely available.  The key specifications are a
> >> 5V to 12V operating voltage and low capacitance.
>

> >>http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=CG06...
> >>http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=ESD0...

>
> >> I have posted this before but you should also look over the following
> >> for notes on USB peripheral design.  Pg.9 and 12 specifically.
> >>http://www.ti.com/sc/docs/apps/msp/intrface/usb/emitest.pdf
>

> >> On Fri, Apr 8, 2011 at 8:22 PM, Bob Paddock<graceindustr...@gmail.com>  wrote:

> >>> On Fri, Apr 8, 2011 at 3:58 PM, L szl Monda<l...@monda.hu>  wrote:
>
> >>>> I wonder what you think about the necessity of USB6B1 part -
> >>>>http://www.st.com/stonline/books/pdf/docs/5605.pdf
>
> >>>> I assume that laptops / PCs should have some similar parts inside of
> >>>> them
> >>> They should.  Some manufactures may skimp to cut costs.  Hard to prove.
>
> >>>> and putting it into a device is overkill.
> >>> Not at all. The device needs protection as well.
>
> >>>>   Anyways, I don't have
> >>>> a clue about analog issues and I'm just guessing so I'm interested
> >>>> about your opinions.
> >>> Static protection should always be part of the USB hardware design.
> >>> I gave an introduction to static issues in my blog:
>

> >>>http://blog.softwaresafety.net/2010/04/software-bug-or-electrostatic-...

>
> >>> --
> >>> You received this message because you are subscribed to the Google Groups "LUFA Library Support List" group.
> >>> To post to this group, send email to lufa-s...@googlegroups.com.
> >>> To unsubscribe from this group, send email to lufa-support...@googlegroups.com.

> >>> For more options, visit this group athttp://groups.google.com/group/lufa-support?hl=en.

>
> >> --
> >> You received this message because you are subscribed to the Google Groups "LUFA Library Support List" group.
> >> To post to this group, send email to lufa-s...@googlegroups.com.
> >> To unsubscribe from this group, send email to lufa-support...@googlegroups.com.
> >> For more options, visit this group athttp://groups.google.com/group/lufa-support?hl=en.
>
> --
> Steven Riedl

> ried...@gmail.com

[László Monda]

Thanks a lot for elaborating on that. I'll consider using some
protection mentioned here.

[Bob Paddock]

On Fri, Apr 8, 2011 at 5:16 PM, László Monda <la...@monda.hu> wrote:
> My question is
> that how much overcurrent / ESD is a problem in the real life and how
> much should I worry about it.  Let's say it affects only 1 people out
> of 10,000 in general in which case I wouldn't worry too much.

Fundamentally comes down to how much is your reputation worth?

That one person with the bad device might post on Facebook, or here or anyplace,
that your products are junk, and then a whole bunch of people won't
buy/use them.

ESD is a real world problem that does not get enough attention.

[Bob Paddock]

> You can use simpler and cheaper bidirectional ESD suppressors.  They
> simplify routing as you only need to connect them between the data
> lines and the connector shield (the bulk of your USB connector).

There can be differential potentials that are not common-mode.
In other words there should be protection between +D and -D, not just
+D and -D to ground.
Vusb also needs protected.

Never rely on the internal ESD protection of an IC to protect a system.

Make sure any protection IC has a low capacitance rated for placing on
the data lines.

[Opendous]

I recently came across some useful notes on USB hardware design:
http://electronics.stackexchange.com/questions/2425/are-common-mode-choke-coils-needed-on-usb

For a more up-to-date review of ESD best practices read Chapter 5 of
"High Speed USB Platform Design Guidelines":
http://www.usb.org/developers/docs/hs_usb_pdg_r1_0.pdf

For the most up-to-date best practices check out the SuperSpeed USB
design guidelines, pg.23 forward:
http://www.usb.org/developers/presentations/pres0311/2-4_SSUSB_DevCon_DesignGuidelines_Heck.pdf

[László Monda]

Thanks for sharing all this info guys, you've persuaded me about the
importance of protecting USB devices.

I think I'll use 2 x STF202-22 for lower-end projects and 1 x
STF202-22 for higher-end projects. I'll never understand the analog
world fully but I think using such parts is a good practice.

> --
> You received this message because you are subscribed to the Google Groups "LUFA Library Support List" group.
> To post to this group, send email to lufa-s...@googlegroups.com.
> To unsubscribe from this group, send email to lufa-support...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/lufa-support?hl=en.
>
>

--
László Monda <http://monda.hu>

[Opendous]

Before using the STF202-22 note that it only mentions USB 1.1
compatibility. However, nothing in the datasheet precludes USB 2.0
Full Speed compatibility.

Also, there is the problem of doubling up the pull-up resistor. Both
the STF202-22 and the USB AVRs have pull-up resistors on the data
lines to signal USB FS or LS operation. You will need to set the
DETACH bit to disable the pull-up. However, I'm not sure if the AVR
will then assume it is operating in Low Speed mode. Test this before
proceeding. The USB AVRs might require a specific sequence of
operations to allow USB.

If you do use the STF202-22 or something similar remember to route the
lines correctly so that the pull-up is on the D+ line for USB Full
Speed operation.

Your safest bet is to simply use two ESD suppressors.

On Apr 13, 6:49 pm, László Monda <l...@monda.hu> wrote:
> Thanks for sharing all this info guys, you've persuaded me about the
> importance of protecting USB devices.
>
> I think I'll use 2 x STF202-22 for lower-end projects and 1 x
> STF202-22 for higher-end projects.  I'll never understand the analog
> world fully but I think using such parts is a good practice.
>
>
>
> On Tue, Apr 12, 2011 at 5:31 PM, Opendous <opendous.supp...@gmail.com> wrote:
> >  I recently came across some useful notes on USB hardware design:

> >http://electronics.stackexchange.com/questions/2425/are-common-mode-c...

>
> >  For a more up-to-date review of ESD best practices read Chapter 5 of
> > "High Speed USB Platform Design Guidelines":
> >http://www.usb.org/developers/docs/hs_usb_pdg_r1_0.pdf
>
> >  For the most up-to-date best practices check out the SuperSpeed USB
> > design guidelines, pg.23 forward:

> >http://www.usb.org/developers/presentations/pres0311/2-4_SSUSB_DevCon...

>
> > On Apr 11, 12:24 pm, Bob Paddock <graceindustr...@gmail.com> wrote:
> >> > You can use simpler and cheaper bidirectional ESD suppressors.  They
> >> > simplify routing as you only need to connect them between the data
> >> > lines and the connector shield (the bulk of your USB connector).
>
> >> There can be differential potentials that are not common-mode.
> >> In other words there should be protection between +D and -D, not just
> >> +D and -D to ground.
> >> Vusb also needs protected.
>
> >> Never rely on the internal ESD protection of an IC to protect a system.
>
> >> Make sure any protection IC has a low capacitance rated for placing on
> >> the data lines.
>
> > --
> > You received this message because you are subscribed to the Google Groups "LUFA Library Support List" group.
> > To post to this group, send email to lufa-s...@googlegroups.com.
> > To unsubscribe from this group, send email to lufa-support...@googlegroups.com.

> > For more options, visit this group athttp://groups.google.com/group/lufa-support?hl=en.

[László Monda]

On Thu, Apr 14, 2011 at 12:01 AM, Opendous <opendous...@gmail.com> wrote:
> Before using the STF202-22 note that it only mentions USB 1.1
> compatibility.  However, nothing in the datasheet precludes USB 2.0
> Full Speed compatibility.

Good catch! Anyone knows about a compatible USB2.0 (or above) part?

> Also, there is the problem of doubling up the pull-up resistor.  Both
> the STF202-22 and the USB AVRs have pull-up resistors on the data
> lines to signal USB FS or LS operation.  You will need to set the
> DETACH bit to disable the pull-up.  However, I'm not sure if the AVR
> will then assume it is operating in Low Speed mode.  Test this before
> proceeding.  The USB AVRs might require a specific sequence of
> operations to allow USB.
>
> If you do use the STF202-22 or something similar remember to route the
> lines correctly so that the pull-up is on the D+ line for USB Full
> Speed operation.

Thanks, I'll keep those advices in mind.

> Your safest bet is to simply use two ESD suppressors.

I pretty much agree after reading all the above.

> For more options, visit this group at http://groups.google.com/group/lufa-support?hl=en.

[László Monda]

Can you guys recommend an alternative part for F2594CT which is 1206
sized? I'm not sure what values should I filter in the Digi-Key
search.

[Opendous]

If you can prototype a 0.8mm pitch TQFP AT90USB/ATmega32U4 then you
can prototype 0603 components (1.5mm pitch). The CG0603MLC-05LECT-ND
is a cheaper option. You cannot use 1206 (3.5mm) size or larger ESD
suppressors as they have very high capacitance and are aimed at power
supplies, not signals. For example, the following 1206 ESD has 200pF
of capacitance. The SOT23 STF202-22 has 68 and the 0603 F2594CT
(PGB1010603) has just 0.06pF.
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=478-2530-1-ND

The USB Compliance checklist Pg.6 M20 requires that they have <75pF:
http://www.usb.org/developers/compliance/check_list/compchkotg1_0080205.pdf

I have successfully used the dual-line PRTR5V0U2X,215 in a couple of
High Speed USB projects which is SOT143B (1.7mm pin spacing) and has
1.5pF capacitance:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=568-4140-1-ND

There is also a SOT23 (2.2mm pitch) version which should be
acceptable (75pF) but is aimed at SIM cards:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=568-4050-1-ND

On Apr 23, 9:22 pm, László Monda <l...@monda.hu> wrote:
> Can you guys recommend an alternative part for F2594CT which is 1206
> sized?  I'm not sure what values should I filter in the Digi-Key
> search.
>
>
>

> On Thu, Apr 14, 2011 at 12:37 AM, László Monda <l...@monda.hu> wrote:

> László Monda <http://monda.hu>

[G Bulmer]

I am very interested in compact USB protection for full-speed USB
electronics.

I want something which does NOT include the data-line pull-up
resistor, so the STF202-22 is not easily applicable.
I want it to be easy to develop and experiment with different USB
devices, so I want to control the Rup resistor from software myself,
and not have it hard-wired on.

I found what looks like the ideal device, NUF2030X6:
https://www.onsemi.com/PowerSolutions/product.do?id=NUF2030X6

It contains the two 22ohm termination resistors, two 30pF capacitors,
and five ESD protectors in a tiny 1.6mm square package.
BUT, I can't find anyone who sells it in small quantity.

It's sibling is the NUF2042XV6:
https://www.onsemi.com/PowerSolutions/product.do?id=NUF2042XV6
the only difference is the capacitors are 42pF.
Do folks think that would be okay? It still loks within USB spec.

This is available, for example Mouser sell it:
http://www2.mouser.com:80/Search/Refine.aspx?Keyword=NUF2042XV6
though others like AVnet.

TIA

On Apr 9, 10:50 pm, Opendous <opendous.supp...@gmail.com> wrote:
> >What about less stricter product categories like
> >consumer products or developer boards?
>
>   You are most likely to encounter IC-damaging situations with
> consumers.  ESD isn't a consumer safety issue, it is a device safety
> issue.  In general you are correct that it is probably a 1 in 10000
> connections issue but heat and dry conditions increase the likelihood
> of ESD events (as do carpeted environments, wearing wool, or other
> static generating situations).  The Arduino UNO uses PGB1010603MR ESD

> Suppressors:http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=F259...

>
> >STF202-22
>
>   I was unaware of such an all-in-one IC and it is indeed a perfect

> solution:http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=STF2...

[G Bulmer]

I stumbled across what looks like an even better USB protection device
than the NUF2042XV6, the Semtech EClamp2522P:
http://www.semtech.com/circuit-protection/emi-filters/eclamp2522p/

It has termination resistors, filter capacitors, ESD protection, *and*
has the 1.5K pull-up on its own pin, so that it can be controlled
externally, rather than be hard-wired. It is also sold by several
suppliers.

Has anyone any experience of it?

[Opendous Support]

If you want general availability and ease of use stick to a pair of
ESD suppressors on the data lines which come as small as 0201 and some
type of surge protector on the VBUS line:
http://en.wikipedia.org/wiki/Surge_protector

TI's "Latch-Up,ESD,And Other Phenomena" Application Report is a good read:
http://focus.ti.com/lit/an/slya014a/slya014a.pdf

>the capacitors are 42pF... It still looks within USB spec.

Yes it is, the USB Compliance checklist Pg.6 M20 requires that they have <75pF:
http://www.usb.org/developers/compliance/check_list/compchkotg1_0080205.pdf

[G Bulmer]

Thank you for the link to TI's "Latch-Up,ESD,And Other Phenomena"
Application Report, I'll read it in the morning

AND8200/D "Design Considerations for ESD/EMI Filters: I (Almost
Everything You Wanted to Know About EMI Filters and Were Afraid to
Ask)"
http://www.onsemi.com/PowerSolutions/supportDoc.do?type=AppNotes&rpn=NUF2042XV6
Was quite interesting.

Thank you for the suggestion on ESD suppressors. I realise this might
sound a bit mad, but I don't want to use very small components.
I am trying to ensure my design can be made by hand, in a 'toaster
oven". So reducing the number of parts is good, but going below 0603
is bad.

I'll give that Semtech EClamp2522P a try.

On May 25, 12:20 am, Opendous Support <opendous.supp...@gmail.com>
wrote:

>   If you want general availability and ease of use stick to a pair of
> ESD suppressors on the data lines which come as small as 0201 and some
> type of surge protector on the VBUS line:http://en.wikipedia.org/wiki/Surge_protector
>
>   TI's "Latch-Up,ESD,And Other Phenomena" Application Report is a good read:http://focus.ti.com/lit/an/slya014a/slya014a.pdf
>
> >the capacitors are 42pF... It still looks within USB spec.
>

>  Yes it is, the USB Compliance checklist Pg.6 M20 requires that they have <75pF:http://www.usb.org/developers/compliance/check_list/compchkotg1_00802...

[Opendous Support]

Thanks for posting that OnSemi AppNote.

>going below 0603 is bad

I have been using the 0603 Stackpole ESD03A5V5R17V recently for all
my designs that require discrete ESD protection. It has very low
capacitance (<0.2pF) and is cheap in moderate quantities. I have also
successfully used the 0603 Bourns CG0603MLC-05LE (<0.5pF) and 0402
CG0402MLC-05LG.

As for VBUS line protection, you can simply use a ferrite bead on
the VBUS line. For more significant protection add a 1206 or 0805 PTC
Fuse. Either way you can reuse the same footprint for either a
ferrite bead or PTC Fuse. Make sure the PTC Fuse has a Hold Current
of at least 500mA.
http://www.bourns.com/data/global/pdfs/Bourns_USB_2_0_PortNote.pdf

>I'll give that Semtech EClamp2522P a try.

Is it stocked anywhere or a sample?

Let us know how things go.

[G Bulmer]

Thank you again :-)

On May 25, 1:02 pm, Opendous Support <opendous.supp...@gmail.com>
wrote:

>   Thanks for posting that OnSemi AppNote.
>
> >going below 0603 is bad
>
>   I have been using the 0603 Stackpole ESD03A5V5R17V recently for all
> my designs that require discrete ESD protection.  It has very low
> capacitance (<0.2pF) and is cheap in moderate quantities.  I have also
> successfully used the 0603 Bourns CG0603MLC-05LE (<0.5pF) and 0402
> CG0402MLC-05LG.

When I said make it easy to make by hand, my target builders are
school children and amateurs.
I find 0603 tiring, so I am worried that is too small for others.

My local schools use 1206 parts to introduce school children to SMD
electronic construction (using the toaster oven process).
My design might be the first that uses smaller components than 1206. I
am trying to stay with 0805 or larger to make it easier for folks.

Since I posted, I realised the EClamp2522P might be too challenging
because it has a pad underneath, so probably back to either the
NUF2042XV6 or normal size resistors, capacitors, and a USB protection
chip like a IP4220CZ6 or something.

>
>   As for VBUS line protection, you can simply use a ferrite bead on
> the VBUS line.  For more significant protection add a 1206 or 0805 PTC
> Fuse.  Either way you can reuse the same footprint for either a
> ferrite bead or PTC Fuse.  Make sure the PTC Fuse has a Hold Current
> of at least 500mA.http://www.bourns.com/data/global/pdfs/Bourns_USB_2_0_PortNote.pdf

Excellent! Thanks for the link.

I have been searching around for something which gives a specific part
value for the VBUS PTC, but didn't find anything.

By luck, it happens that I had chosen MF-NSMF075 on the basis that it
is the cheapest part, with a holding current above 500mA, that I could
find at RS :-), but I was (literally now) about to order several
different ones, and do some experiments.

So that application note is a time saver, money saver, and good
reassurance.

>
> >I'll give that Semtech EClamp2522P a try.
>
>   Is it stocked anywhere or a sample?

Yes, RS have it in stock. Future list it, but have no stock. But, I
think the footprint will make it too hard for children to solder.

>
>   Let us know how things go.

Will do, and thanks again for your help.

Do you have a simple process/test rig you use for testing the
protection circuitry?

>
>
>
> On Tue, May 24, 2011 at 11:37 PM, G Bulmer <gbul...@gmail.com> wrote:
> > Thank you for the link to  TI's "Latch-Up,ESD,And Other Phenomena"
> > Application Report, I'll read it in the morning
>
> > AND8200/D "Design Considerations for ESD/EMI Filters: I (Almost
> > Everything You Wanted to Know About EMI Filters and Were Afraid to
> > Ask)"

> >http://www.onsemi.com/PowerSolutions/supportDoc.do?type=AppNotes&rpn=...

> ...
>
> read more »

[László Monda]

I'm also interested about how to test the protection circuitry.

As a related issue, are there any other components that are suggested
on USB devices? For example I've seen 500mA fuses on some of such
devices?

As a final question are there any rules regarding the placement of
various components? I know that the crystal has to be placed as close
to the MCU as possible and I think (not sure) that the suppressors
have to be placed near to the USB connector.

[Bob Paddock]

On Thu, May 26, 2011 at 6:36 PM, László Monda <la...@monda.hu> wrote:

> As a final question are there any rules regarding the placement of
> various components?

High current transients, like a lightning strike, will take the path
of least inductance not the path of least resistances as is commonly
believed. Adding vias and sharp right angle turns in the traces would
be a good thing from the perspective of a lightning hit, and a bad
idea in just about every other regard. My point here is that you
really need to know what you are protecting against to do the best
layout.

>I think (not sure) that the suppressors
> have to be placed near to the USB connector.

Put the suppressors as close to the point of ingress as possible.
They lose effectiveness the farther they are from the ingress point.
Think of it as the farther away they are, the more chance the
transient has to do damage to more stuff because of more exposure
(longer traces would be closer to more stuff).

[László Monda]

Thanks a lot, Bob. Very useful insights.

[László Monda]

Looks like this thread gets resurrected from time to time. I have one
question left.

I've observed that various boards, such as Adafruit's Atmega32u4
Breakout Board+ and Arduino boards use a 500mA fuse. In what scenario
is this useful? Is it a good idea to use it even if I use ESD
suppressors?

Thanks!

[Opendous Support]

The fuse and ESD suppressors solve separate issues. The fuse
protects against short circuits and overcurrent conditions.

Arduino Uno uses a very large PTC Resettable Fuses with 550mA Hold
Current (Ih):
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=MF-SMDF050-2CT-ND

However, such fuses derate to Ih=230mA at +85C.
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=507-1520-1-ND

There are smaller and cheaper fuses with Ih=750mA which derate to
Ih=460mA at +85C. Use a trace calculator to make sure your design can
handle the trip current at your target ambient temperature.
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=507-1481-1-ND
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=507-1520-1-ND
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=MF-PSMF075X-2CT-ND

[László Monda]

Thanks a lot Matt, you're the undisputed champion of USB analog issues!

You've just made me aware of PTC fuses and went from zero to hero
(well, not quite yet) in a short while.

In overall I like the 0ZCA0075FF2G quite a lot. I guess it should be
pretty good unless I'm the middle of the Sahara with a high powered
USB device in which case I might just be screwed.

I should place the fuse directly next to the VBUS pin of the USB socket, right?

Are these devices polarized by any chance?

[Opendous Support]

No, not polarized. Just place the PTC Fuse in series with the VBUS
line and as close as possible to the USB Connector.

[László Monda]

Sounds like a walk in the park. Thanks a lot!

[László Monda]

Lots of time has been spent since my last post, but let me resurrect
this thread because I'm having a related issue.

I've recently finished soldering my custom designed PCB which features
the following USB circuit:

* USB connector D+ lead -- ESD03A5V5R17V -- 22ohm resistor -- ATmega32U2 D+ pin
* USB connector D- lead -- ESD03A5V5R17V -- 22ohm resistor -- ATmega32U2 D- pin

Turns out my design works only without the ESD03A5V5R17V parts.

Anyone knows why?

Thanks in advance!

On Mon, Apr 25, 2011 at 12:31 AM, László Monda <la...@monda.hu> wrote:
> Thanks a lot for all the useful informatiton on the subject!  I'm
> grateful beyond measure.
>
> I'll go with ESD03A5V5R17V, too.
>
> By the way, didn't you forget to CC the LUFA mailing list?
>
> On Sun, Apr 24, 2011 at 7:18 PM, Opendous <opendous...@gmail.com> wrote:
>> Check out my Micropendous-32U2 project:
>> http://code.google.com/p/micropendous/wiki/Micropendous_32U2
>>
>> I use the 0603-size ESD03A5V5R17V (little blue parts next to USB-miniB
>> connector) which are the cheapest ESD Suppressors when buying 5000-
>> piece reels and have <0.02pF of capacitance.  All the other parts on
>> that board are 0805 size.  Notice there isn't that much of a size
>> difference.  0402 are significantly smaller.
>>
>> I have a Bill Of Materials available:
>> http://code.google.com/p/micropendous/source/browse/trunk/Micropendous/Design/Micropendous-32U2/Micropendous-32U2_BillOfMaterials.txt

>>
>> On Apr 23, 9:22 pm, László Monda <l...@monda.hu> wrote:
>>> Can you guys recommend an alternative part for F2594CT which is 1206
>>> sized?  I'm not sure what values should I filter in the Digi-Key
>>> search.
>>>
>>>
>>>

>>> On Thu, Apr 14, 2011 at 12:37 AM, László Monda <l...@monda.hu> wrote:

>>> László Monda <http://monda.hu>
>>
>
>
>
> --
> László Monda <http://monda.hu>
>

--
László Monda <http://monda.hu>

[Opendous Support]

Are the ESD03A5V5R17V in series with the USB data lines? The ESD
protection should shunt to the USB connector's shield.

If your circuit is correct then please post a picture.

[László Monda]

You've made me realize that I've totally done this the wrong way. I
thought that I have to flow D+ / D- *through* the suppressors. Ouch!

On Wed, Oct 26, 2011 at 6:23 PM, Opendous Support
<opendous...@gmail.com> wrote:
> Are the ESD03A5V5R17V in series with the USB data lines?  The ESD
> protection should shunt to the USB connector's shield.

What do you mean by the "shield"? Do you mean the outer shell of the
connector, so to speak? If so, I should route the shield signals from
the feet of the connector (that also provide mechanical stability)
right?

> For more options, visit this group at http://groups.google.com/group/lufa-support?hl=en.

[Opendous Support]

Yes, the USB lines connect to the USB connector bulk through the ESD device(s).

[László Monda]

First of all sorry for the long delay. (The LUFA mailing list is
filtered into a subfolder of mine which I couldn't see right away and
your answer wasn't CC'ed so it didn't landed in my inbox immediately.)

Thanks a lot for the explanation, that made things a lot clearer to
me. I still have some questions.

1) There's a device between VCC and GND in your schematic that is
labelled with "F0" and "L1" (I'm not so sure about "F0" because the
text is rather low resolution). It looks like an inductor. Which is
this specific part and why is it needed?

2) As for the other parts of the circuit, such as the GND pins of the
MCU, I should use GND and not GNDPWR (the shield) to connect those,
right? I suppose GNDPWR is only for protection purposes.

3) What is / are the EDA application(s) that have been used for
creating USB_Connection_Schematic.jpg and USB_ESD_Layout.jpg? Those
pictures look very good, way better than Eagle (which I use).

Thanks in advance!

[László Monda]

On Thu, Nov 3, 2011 at 8:25 PM, Opendous Support
<opendous...@gmail.com> wrote:
>  Grab a copy of KiCAD and view the full design at full resolution:
> http://iut-tice.ujf-grenoble.fr/cao/
>
>  If using the GoogleCode web interface to download the files instead
> of SVN you will at least need '.brd' '.sch' and '.lib'  and will have
> to click on each file then right-click on "View raw file" and "Save
> As...".
> http://code.google.com/p/micropendous/source/browse/trunk/Micropendous/Design/Micropendous-A/

I was well aware that KiCAD exists, is getting more and more evolved
and was thinking about learning it and eventually migrating to it but
I'd have never expected the quality of rendering that I've just seen
in your attached images. I'm definitely gonna give it a try in the
near future.

By the way, what's your opinion about gEDA? Considering that it seems
a lot less popular than KiCAD suggests me that its usability must be
inferior.

>>1) There's a device between VCC and GND in your
>>schematic that is labelled with "F0" and "L1"
>

>  There is nothing between VCC and GND.

Ah, I meant GNDPWR and GND, sorry.

>  There is a PTC fuse (F1)
> between the USB connector's VBUS and the board's VBUS:
> http://search.digikey.com/us/en/products/0ZCE0075FF2E/507-1520-1-ND

I could recognize that it's supposed to be a PTC fuse. I'm using
http://search.digikey.com/us/en/products/0ZCA0075FF2G/507-1481-1-ND/2165310
successfully so far.

>  There is a ferrite bead (L1) connecting SHIELD (GNDPWR) and GND:
> http://search.digikey.com/us/en/products/HZ0805E601R-10/240-2399-1-ND

I'm adding this to my repertoire of USB support circuitry components.
Is this supposed to reduce noise?

>>2) As for the other parts of the circuit, such as
>>the GND pins of the MCU, I should use GND
>

>  Correct.  GNDPWR is the noisy cable shield.  GND is the USB GND.
>
> 3) What is / are the EDA application(s) that [you used]
>
>  KiCAD.  It is completely Free and Open Source Software with free and
> Open file formats:
> http://www.lis.inpg.fr/realise_au_lis/kicad/

Thanks a lot! You've taught me so much so far.