[SI-LIST] Re: Guard Traces

  • From: steve weir <weirsi@xxxxxxxxxx>
  • To: si-list@xxxxxxxxxxxxx
  • Date: Tue, 20 Mar 2012 03:53:53 -0700

On 3/20/2012 2:13 AM, Rohit MISHRA wrote:
> Thanks Wolfgang,
>
> Your mail made me think that if there is some problem in my understanding 
> about impedance of line(trace).
>
> I understand that a uncoupled line always have some characteristic impedance 
> and now If I bring other trace closer to uncoupled line, their fringe 
> electric and magnetic fields will overlap and the coupling between them will 
> increase. But capacitive&  inductive coupling between traces doesn't always 
> mean that my first trace will see different impedance. It's potential at 
> second trace that decides if total field seen by first trace has changed, 
> right ?
Wrong.  Impedance is just the transfer function of V/I.  When you 
introduce coupling of any kind you alter I vs V and therefore the 
impedance.  The only issue is by how much at a given frequency.
>
> My understanding is that when my second trace is at ground potential( like in 
> guard trace), my first trace should still see almost same field which it was 
> seeing when it was uncoupled so its impedance shouldn't change.
>
> In case when second trace is driven with differential or common mode signal( 
> same potential as first trace ) then I can understand that total field seen 
> by first trace will change and hence change in its impedance.
>
> Do u see any problem in my point ?
>
> Rohit
>
>
>
>
> -----Original Message-----
> From: Wolfgang.Maichen@xxxxxxxxxxxx [mailto:Wolfgang.Maichen@xxxxxxxxxxxx]
> Sent: Tuesday, March 20, 2012 1:33 PM
> To: Rohit MISHRA; fred@xxxxxxxxxxxxx; si-list@xxxxxxxxxxxxx
> Subject: RE: [SI-LIST] Re: Guard Traces
>
> In theory a guard trace will ALWAYS affect the line impedance (it lowers the 
> impedance). The question is just how much. The farther it gets away, the less 
> effect it has, but there is no distance where the effect is exactly zero. As 
> a rough estimate, once the guard trace is more than about 3 line widths away 
> the effect will be negligible. But don't take my word for it, a 2D field 
> solver (e.g. TNT/MMTL) can give you precise answers. In any case, a guard 
> trace at that distance will neither help not hurt with crosstalk or signal 
> integrity.
>
> Wolfgang
>
>
>
> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On 
> Behalf Of Rohit MISHRA
> Sent: Tuesday, March 20, 2012 8:53 AM
> To: fred@xxxxxxxxxxxxx; si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: Guard Traces
>
> Fred,
>
> Today, I was going through your below mail and in this mail, you said that as 
> the guard trace gets closer, the impedance of the line changes. I want to 
> know that when u say closer, do u mean to say guard trace closer than the 
> line and ground plane separation.
>
> My understanding says that if separation between guard trace and the line is 
> more than line and plane separation than ground potential at guard trace 
> should not affect current through the line and hence no change in line 
> impedance.
>
>
> Rgds,
> Rohit
>
>
> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On 
> Behalf Of Fred Balistreri
> Sent: Saturday, March 10, 2012 6:12 AM
> To: doh@xxxxxxxxxx; weirsi@xxxxxxxxxx; si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: Guard Traces
>
> As the guard trace gets closer the impedance of the line changes. So you'll
> have to use a
> Field solver to predict.  If you can get it close and meet Zo then you're
> ok.   If it's high
> Speed then stiching vias to the gnd plane may be required.  Richie's point
> is that you usually
> can get the noise immunity you're looking for by simply using proper spacing
> and it may take up less geometry with less fuss.
>
> Now there are cases where guard traces are useful. If one doesn't have
> planes for example or
> some analog applications where sensitive lines need to be carefully isolated
> from digital, switching supply or other such sources.
>
> Bottom line depends on the application and skill set of the
> designer/engineer.
>
> Cheers
>
> Fred
>
> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
> Behalf Of Oh, Dan
> Sent: Friday, March 09, 2012 4:14 PM
> To: weirsi@xxxxxxxxxx; si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: Guard Traces
>
>
> I agree completely with Steve.
> If the spacing between signal and guard trace is sufficiently narrow than
> dielectric height, it will help. It is all about assumption and we should
> not simply discount the guard trace for all cases. Cheers,
>
> -Dan Oh
>
> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
> Behalf Of steve weir
> Sent: Friday, March 09, 2012 3:21 PM
> To: si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: Guard Traces
>
> Rick, the percentage coupling between adjacent lines depends on the
> geometry.  For boards with modest dielectric thickness above solid planes
> the percentage coupling is small.  As the dielectric gets thicker, the
> percentage coupling between adjacent lines goes up.
>
> Steve.
>
> On 3/9/2012 8:48 AM, Rick Collins wrote:
>> I hate posted to this group because of all the Out Of Office replies,
>> but I can't let this go.
>>
>> It is exactly the inductive coupling that a guard trace has no impact
>> on, right?  The concept of the guard trace is that it is closer to the
>> victim trace than the aggressor is and so the electrostatic potential
>> of the guard is the dominant influence.  But the inductive coupling is
>> not significantly affected.  I supposed there could be some return
>> current induced into the guard trace, but this would be a small effect
>> relative to the ground and power planes.
>>
>> Do I have this all wrong?  Does a guard trace reduce inductive
>> coupling via the induced ground current?
>>
>> Rick
>>
>>
>> At 11:40 AM 3/9/2012, Oh, Dan wrote:
>>> Hi Steve and Doug,
>>> I also agree with you. You guys have one more vote. The guard trace
>>> is very effective to reduce the inductive coupling which cannot be
>>> easily reduced by spacing.
>>> Best,
>>> -Dan Oh
>>> -----Original Message-----
>>> From: si-list-bounce@xxxxxxxxxxxxx
>>> [mailto:si-list-bounce@xxxxxxxxxxxxx] On Behalf Of steve weir
>>> Sent: Thursday, March 08, 2012 9:23 PM
>>> To: si-list@xxxxxxxxxxxxx
>>> Subject: [SI-LIST] Re: Guard Traces
>>>
>>> Doug, I agree that guard traces can help in certain situations.  In
>>> most, and particularly when inserted without doing the requisite
>>> engineering work, they can make a good mess of things.
>>>
>>> Best Regards,
>>>
>>>
>>> Steve.
>>> On 3/8/2012 8:22 PM, Doug Smith wrote:
>>>> Hi Lee,
>>>>
>>>> Your statement is true for most cases, but for two layer boards
>>>> (every household has several of these with processors on them) guard
>>>> traces can stabilize impedance, improve grounding, and lower noise
> coupling.
>>>> I agree with you for advanced electronics.
>>>>
>>>> We should always state the assumptions made when citing design or
>>>> other rules. I see this a lot in EMC books where an equation is
>>>> given (such as slot radiation) but the (sometimes many) assumptions
>>>> are not given and people use the equations when they don't apply.
>>>> Shielding effectiveness is another case where the common equations
>>>> do not apply in many cases (far field conditions are assumed when in
>>>> fact they are not met). Under every equation in EMC and other books
>>>> there should be a list of assumptions given.
>>>>
>>>> Doug
>>>>
>>>> On 3/8/12 10:52 AM, Lee Ritchey wrote:
>>>>> With all the feedback on this forum that guard traces are a bad
>>> idea, why is this discussion still going on?  Are there still people
>>> who cling to the notion that guard traces are useful, even in the
>>> face of mounting evidence that they don't help?
>>>>> -----Original Message-----
>>>>>> From: Dennis<dennis.han@xxxxxxxxx>
>>>>>> Sent: Mar 7, 2012 12:59 PM
>>>>>> To: si-list@xxxxxxxxxxxxx
>>>>>> Subject: [SI-LIST] Re: Guard Traces
>>>>>>
>>>>>> In addition to what Wolfgang has stated, if the guard traces
>>> are too close to the signal traces, and most people who believe in
>>> ground traces put them too close (ground fill also falls in this
>>> category), the impedance of the signal traces is lowered and there
>>> are less crosstalk and radiation because of it if the now mismatched
>>> terminations don't cause other problems.
>>>>>>
>>>>>>
>>>>>>
>>>>>> --- In si-list@xxxxxxxxxxxxxxx, Gwen and Wolfgang<gwmaichen@...>
> wrote:
>>>>>>> Hello,
>>>>>>>
>>>>>>> first, the short answer: simply don't use guard traces here, they
>>>>>>> won't help much with isolation of high speed signals, and they
>>>>>>> use up way too much routing space (on all layers, because of the
>>>>>>> stitching vias necessary). Much easier to just increase the
>>>>>>> distance between diff pairs etc. and you get the same benefit.
>>>>>>> (in fact, I suspect the reason many people believe in the magic
>>>>>>> power of guard traces in this setting - high speed signaling - is
>>>>>>> because those traces _force_ you to have enough space between
>>>>>>> signal lines...). Guard traces have their place in low-current /
>>>>>>> low-leakage / high-voltage applications, but NOT in high speed
>>> digital. At least that's my opinion.
>>>>>>> If you do need grounded traces, stitch them to the ground planes
>>>>>>> with vias spaced no further than a quarter wavelength from the
>>>>>>> signal trace and at an interval of no less than a quarter wave
>>>>>>> length from each other. The wavelength given of course by
>>>>>>> bandwidth connected with the signal rise time, BW = 0.33 / Tr, or
>>>>>>> a bit more conservatively, the knee frequency f_knee = 0.5 / Tr.
>>>>>>> No special termination needed at the ends of those traces because
>>>>>>> they are already solidly grounded along their full length.
>>>>>>>
>>>>>>> Wolfgang
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> On 3/7/2012 7:31 PM, Aaditya Kandibanda wrote:
>>>>>>>> Hello Everyone,
>>>>>>>> I have few questions about the guard traces.
>>>>>>>>
>>>>>>>> 1. How do I terminate the guard traces?
>>>>>>>> 2. While using the guard traces for differential pair, how shall
>>>>>>>> I place them between the differential pair?
>>>>>>>> 3. How should I place the vias for guard traces?
>>>>>>>> 4. What are disadvantages of Guard traces?
>>>>>>>>
>>>>>>>> Thanks and regards
>>>>>>>> Aaditya
>>>>>>>>
>>>>>>>>
>>>>>>>> ----------------------------------------------------------------
>>>>>>>> --
>>>>>>>> To unsubscribe from si-list:
>>>>>>>> si-list-request@... with 'unsubscribe' in the Subject field
>>>>>>>>
>>>>>>>> or to administer your membership from a web page, go to:
>>>>>>>> http://www.freelists.org/webpage/si-list
>>>>>>>>
>>>>>>>> For help:
>>>>>>>> si-list-request@... with 'help' in the Subject field
>>>>>>>>
>>>>>>>>
>>>>>>>> List forum  is accessible at:
>>>>>>>>                     http://tech.groups.yahoo.com/group/si-list
>>>>>>>>
>>>>>>>> List archives are viewable at:
>>>>>>>>             http://www.freelists.org/archives/si-list
>>>>>>>>
>>>>>>>> Old (prior to June 6, 2001) list archives are viewable at:
>>>>>>>>                     http://www.qsl.net/wb6tpu
>>>>>>>>
>>>>>>> -----------------------------------------------------------------
>>>>>>> -
>>>>>>> To unsubscribe from si-list:
>>>>>>> si-list-request@... with 'unsubscribe' in the Subject field
>>>>>>>
>>>>>>> or to administer your membership from a web page, go to:
>>>>>>> http://www.freelists.org/webpage/si-list
>>>>>>>
>>>>>>> For help:
>>>>>>> si-list-request@... with 'help' in the Subject field
>>>>>>>
>>>>>>>
>>>>>>> List forum  is accessible at:
>>>>>>>                    http://tech.groups.yahoo.com/group/si-list
>>>>>>>
>>>>>>> List archives are viewable at:
>>>>>>>              http://www.freelists.org/archives/si-list
>>>>>>>
>>>>>>> Old (prior to June 6, 2001) list archives are viewable at:
>>>>>>>                      http://www.qsl.net/wb6tpu
>>>>>>>
>>>>>> ------------------------------------------------------------------
>>>>>> To unsubscribe from si-list: si-list-request@xxxxxxxxxxxxx with
>>>>>> 'unsubscribe' in the Subject field
>>>>>>
>>>>>> or to administer your membership from a web page, go to:
>>>>>> http://www.freelists.org/webpage/si-list
>>>>>>
>>>>>> For help:
>>>>>> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>>>>>>
>>>>>>
>>>>>> List forum  is accessible at:
>>>>>>                   http://tech.groups.yahoo.com/group/si-list
>>>>>>
>>>>>> List archives are viewable at:
>>>>>>               http://www.freelists.org/archives/si-list
>>>>>>
>>>>>> Old (prior to June 6, 2001) list archives are viewable at:
>>>>>>               http://www.qsl.net/wb6tpu
>>>>>>
>>>>>>
>>>>> I just used the energy it took to get mad and wrote some blues.
>>>>> Count Basie
>>>>>
>>>>> Lee W. Ritchey
>>>>> Speeding Edge
>>>>> P.O. Box 2194
>>>>> Glen Ellen, CA
>>>>> 707-568-3983
>>>>> ------------------------------------------------------------------
>>>>> To unsubscribe from si-list:
>>>>> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject
>>>>> field
>>>>>
>>>>> or to administer your membership from a web page, go to:
>>>>> http://www.freelists.org/webpage/si-list
>>>>>
>>>>> For help:
>>>>> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>>>>>
>>>>>
>>>>> List forum  is accessible at:
>>>>>                    http://tech.groups.yahoo.com/group/si-list
>>>>>
>>>>> List archives are viewable at:
>>>>>                http://www.freelists.org/archives/si-list
>>>>>
>>>>> Old (prior to June 6, 2001) list archives are viewable at:
>>>>>                http://www.qsl.net/wb6tpu
>>>>>
>>>>>
>>>>>
>>> --
>>> Steve Weir
>>> IPBLOX, LLC
>>> 150 N. Center St. #211
>>> Reno, NV  89501
>>> www.ipblox.com
>>>
>>> (775) 299-4236 Business
>>> (866) 675-4630 Toll-free
>>> (707) 780-1951 Fax
>>>
>>> All contents Copyright (c)2012 IPBLOX, LLC.  All Rights Reserved.
>>> This e-mail may contain confidential material. If you are not the
>>> intended recipient, please destroy all records and notify the sender.
>>>
>>> ------------------------------------------------------------------
>>> To unsubscribe from si-list:
>>> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>>>
>>> or to administer your membership from a web page, go to:
>>> http://www.freelists.org/webpage/si-list
>>>
>>> For help:
>>> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>>>
>>>
>>> List forum  is accessible at:
>>>                  http://tech.groups.yahoo.com/group/si-list
>>>
>>> List archives are viewable at:
>>>                   http://www.freelists.org/archives/si-list
>>>
>>> Old (prior to June 6, 2001) list archives are viewable at:
>>>                   http://www.qsl.net/wb6tpu
>>>
>>>
>>> ------------------------------------------------------------------
>>> To unsubscribe from si-list:
>>> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>>>
>>> or to administer your membership from a web page, go to:
>>> http://www.freelists.org/webpage/si-list
>>>
>>> For help:
>>> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>>>
>>>
>>> List forum  is accessible at:
>>>                  http://tech.groups.yahoo.com/group/si-list
>>>
>>> List archives are viewable at:
>>>                   http://www.freelists.org/archives/si-list
>>>
>>> Old (prior to June 6, 2001) list archives are viewable at:
>>>                   http://www.qsl.net/wb6tpu
>>>
>> ------------------------------------------------------------------
>> To unsubscribe from si-list:
>> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>>
>> or to administer your membership from a web page, go to:
>> http://www.freelists.org/webpage/si-list
>>
>> For help:
>> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>>
>>
>> List forum  is accessible at:
>>                  http://tech.groups.yahoo.com/group/si-list
>>
>> List archives are viewable at:
>>                http://www.freelists.org/archives/si-list
>>
>> Old (prior to June 6, 2001) list archives are viewable at:
>>                http://www.qsl.net/wb6tpu
>>
>>
>>
>
> --
> Steve Weir
> IPBLOX, LLC
> 150 N. Center St. #211
> Reno, NV  89501
> www.ipblox.com
>
> (775) 299-4236 Business
> (866) 675-4630 Toll-free
> (707) 780-1951 Fax
>
> All contents Copyright (c)2012 IPBLOX, LLC.  All Rights Reserved. This
> e-mail may contain confidential material. If you are not the intended
> recipient, please destroy all records and notify the sender.
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> http://www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>
> List forum  is accessible at:
>                 http://tech.groups.yahoo.com/group/si-list
>
> List archives are viewable at:
>                  http://www.freelists.org/archives/si-list
>
> Old (prior to June 6, 2001) list archives are viewable at:
>                  http://www.qsl.net/wb6tpu
>
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> http://www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>
> List forum  is accessible at:
>                 http://tech.groups.yahoo.com/group/si-list
>
> List archives are viewable at:
>                  http://www.freelists.org/archives/si-list
>
> Old (prior to June 6, 2001) list archives are viewable at:
>                  http://www.qsl.net/wb6tpu
>
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> http://www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>
> List forum  is accessible at:
>                 http://tech.groups.yahoo.com/group/si-list
>
> List archives are viewable at:
>                  http://www.freelists.org/archives/si-list
>
> Old (prior to June 6, 2001) list archives are viewable at:
>                  http://www.qsl.net/wb6tpu
>
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> http://www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>
> List forum  is accessible at:
>                 http://tech.groups.yahoo.com/group/si-list
>
> List archives are viewable at:
>                  http://www.freelists.org/archives/si-list
>
> Old (prior to June 6, 2001) list archives are viewable at:
>                  http://www.qsl.net/wb6tpu
>
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> http://www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>
> List forum  is accessible at:
>                 http://tech.groups.yahoo.com/group/si-list
>
> List archives are viewable at:
>               http://www.freelists.org/archives/si-list
>
> Old (prior to June 6, 2001) list archives are viewable at:
>               http://www.qsl.net/wb6tpu
>
>
>


-- 
Steve Weir
IPBLOX, LLC
150 N. Center St. #211
Reno, NV  89501
www.ipblox.com

(775) 299-4236 Business
(866) 675-4630 Toll-free
(707) 780-1951 Fax

All contents Copyright (c)2012 IPBLOX, LLC.  All Rights Reserved.
This e-mail may contain confidential material.
If you are not the intended recipient, please destroy all records
and notify the sender.

------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
http://www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field


List forum  is accessible at:
               http://tech.groups.yahoo.com/group/si-list

List archives are viewable at:     
                http://www.freelists.org/archives/si-list
 
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu
  

Other related posts: