Home  Patents  Binary Sampler  Articles  Contact

Introduction

DON'T CUT THE GROUND PLANE. This advice is shouted on every electronic design forum on the web. And for good reason. Cutting the ground plane disrupts the noise currents that normally flow on the outside edge due to skin effect, and may actually increase noise coupled into critical circuits.

The problem of noise and crosstalk is particularly difficult when prototyping high frequency circuits. It may not be desirable to go directly into pcb layout until the grounding issues are completely understood. But conventional prototyping methods make it difficult to get a clear picture of what is needed to obtain the lowest noise.

I plan to give a more detailed description of this method with photos, but in the meantime here is a brief summary I sent in an email:

  One last  point is the input of the 74AC04 will be sensitive  to any
  stray noise  from  the DDS clock and other  logic  switching  in the
  vicinity.

  I minimize  this by mounting the 74AC04 on a separate  platform made
  from a small piece of double-sided copperclad about 0.020" thick.

  All the bypass caps are grounded to the top layer of the copperclad,
  along with the ground for the 74AC04. A single ground goes  from the
  copperclad to  the rest of the system ground. I  try  several points
  around the periphery to select one that gives the lowest crosstalk.

  This keeps  all the input signals on their own reference  plane, and
  minimizes unwanted signals from climbing up onto the top plane where
  they can cause interference and crosstalk.

  When you are making a prototype, the bottom layer of  the copperclad
  can be soldered directly to your regular ground plane. This  makes a
  very solid mounting for the parts, but you can unsolder the  top pcb
  and move  it if necessary to give room for other circuitry. I  put a
  45 degree bevel on the copperclad to provide a bit of relief  so the
  solder doesn't short the top side to ground.

  This method gives the best performance of any that I have tried, and
  works well  into the microwave frequencies, up to the  limit  of FR4
  (about 3GHz).

In The News

Mon Jan 05, 2009: Keven Wheatley, M0KHZ, posted some nice words about this page in his blog:

M0KHZ's Blog

Dedicated to the world of homebrew amateur radio.

Possible improvements to current H mode mixer performance

January 5th, 2009

This is a short blog update, pointing the reader to visit Mike Monett's, Precision System Technologies, web site.

Following the announcement from Martin - PA3AKE regarding the AM noise issue from the current top end DDS chips from Analog Devices, Mike contacted me with some very interesting and helpful information. Mike has developed a couple of techniques to improve the output of a 74AC04 hex inverter, commonly used to convert sine wave to square wave in SDR's and H mode mixers. For the experimenter I highly recommend following this link:

http://pstca.com/spice/74ac04/limiter.htm

Mike has also investigated the effects of power supply ripple on the above inverter, this also very good reading, once again I highly recommend following this link:

http://pstca.com/spice/ripple/ripple.htm

And to finish this update, I'll point you to another page from Mike's site regarding separation of ground planes around switching logic.

http://pstca.com/spice/pstpp/pstpp.htm

All of the above information will be of interest to the experimenter, I believe this information may also lead to increased performance from the current `state of the art' H mode mixer.

End of update.

http://www.m0khz.com/

Apr 16, 2010 Kevin has taken his blog offline, but google still has a reference to his page. Here's the result obtained while searching google for this url: http://pstca.com/spice/pstpp/pstpp.htm

  M0KHZ's Blog » Ham  interest http://pstca.com/spice/pstpp/pstpp.htm.
  All of  the   above   information   will   be   of  interest  to the
  experimenter, I believe this information may also lead to ...

  http://www.m0khz.com/?cat=7paged=2

Fri, 19 Jun 2009: I posted a link to this page in the Time-Nuts forum. Here's a partial copy:

I'm seriously considering making the 53310A DC power section into a switcher. I have developed some techniques that greatly reduce crosstalk by mounting circuits on their own grounded platform. These have a minimum number of connections to the main system ground, and are chosen to minimize switching currents from flowing on the main system ground. I put a brief description of the PST Prototype Platform on my web site at

http://pstca.com/spice/pstpp/pstpp.htm

Another technique that can help is adding an active filter to the output of the power supply. These can be very effective, but there are a number of approaches that are well-publicized that don't work very well. There is more analysis here:

http://pstca.com/spice/ripple/ripple.htm

Finally, Linear Technology supplies Low Noise Switching Regulators that can give ripple values of less than 100uV, which is comparable to or less than many linear regulators. If interested, see the links in the above url.

Noise can be a tricky problem, but the above information might give some ideas to help tame it.

Mike

http://www.mail-archive.com/time-nuts@febo.com/msg20382.html

My post also appears in Brooke Clarke's reply in http://www.mail-archive.com/time-nuts@febo.com/msg20384.html

Mon, 27 Jul 2009 Dr. Eric Bogatin, of Bogatin Enterprises, asked for suggestions to his list of Common Sense Signal Integrity Principles. I sent him an email describing my technique, and here's his reply:

Received: from ironport-out.teksavvy.com ([206.248.154.186])
        by mail.pppoe.ca (Internet Mail Server v1.0) with SMTP id IWH17236
        for ; Mon, 27 Jul 2009 11:06:36 -0400
From: "Eric Bogatin" 
To: "'Mike Monett'" 
Subject: RE: blog post: common sense signal integrity
Date: Mon, 27 Jul 2009 10:06:31 -0500
Organization: Bogatin Enterprises
Thread-Index: AcoOyZJYJ8AYXNoPTpWGtN08Xm9a3QAAV20w
In-Reply-To: <20090727145024.21784466EC1@mx13.bluetie.com>
Message-Id: <20090727150629.172E22DB006@emta3.nyc1.bluetie.com>

-----Original Message-----
From: Mike Monett [mailto:xxxxxx@xxxxxx.com]
Sent: Monday, July 27, 2009 8:08 AM
To: Eric Bogatin
Subject: Re: blog post: common sense signal integrity

  >> Hi-

  >> I thought you might be interested in my recent blog posting:

  >> Common       Sense      Signal     Integrity      Principles:   a
  >> <http://bethesignal.net/blog/>  Baker's   Dozen.  If  you  have a
  >> favorite common  sense principle I left off the list,  drop  me a
  >> note or post a comment.

  >> eric

  >>Dr Eric Bogatin, Signal Integrity Evangelist
  >>Bogatin Enterprises
  >>Setting the Standard for Signal Integrity Training

  > Hi Eric,

  > I want to thank you for the invaluable information you  provide in
  > signal integrity. I learn a lot from your efforts.

  > One trick I have learned to minimize crosstalk is to  put critical
  > circuits on their own section of ground plane. This helps  to keep
  > switching transients local.

  > The sub-ground  is  connected to the main ground  plane  at places
  > that minimize  crosstalk.  Ideally,  high-speed  data transmission
  > lines are  be  differential   to  minimize  crosstalk  from ground
  > currents, but  single-ended  lines can be placed  near  the common
  > ground point.

  > This is also very useful when prototyping. A brief note describing
  > the PST Prototype Platform is at

  > http://pstca.com/spice/pstpp/pstpp.htm

  > Regards,

  > Mike Monett

Mike-

Thanks for the tip!

--eric

*******************************************************
Dr. Eric Bogatin, Signal Integrity Evangelist
Bogatin Enterprises
Setting the Standard for Signal Integrity Training

www.BeTheSignal.com
Blog: www.beTheSignal.com/blog
26235 W 110th Terr
Olathe, KS  66061
e: xxxxxx@xxxxxx.com
v: 913-393-1305   cell: 913-424-4333   f: 913-393-0929
*******************************************************

Tue, 08 Sep 2009: I posted a link to this page in the Time-Nuts forum. Here's a partial copy:

Re: [time-nuts] HP 5071A Electron Multiplier of Cesium Beam Tube

Tue, 08 Sep 2009 11:20:48 -0700

[...]

So the question is what kind of tweaking is needed to get the best performance in a regenerative divider, and what kind of equipment is needed to do it? Then, is perfection really needed in order to beat the Wenzel ULN? Maybe put up with lower performance in the beginning, then upgrade later.

BTW, I love your construction techniques. Everything on ground plane in shielded boxes, such as

http://www.thegleam.com/ke5fx/equinox/equinox.html

One trick I have found that really helps isolate circuit blocks is to put them on their own small island pcb, which is then soldered to the main ground plane to hold it in place. Then find the location of ground connections that give the lowest crosstalk. A brief description is here.

http://pstca.com/spice/pstpp/pstpp.htm

I need to post some pictures to show how easy it is to mount descrete components like caps, transistors and ics. Hopefully I may find the time to do that soon. But this is now my preferred method of construction, both for convenience and very low crosstalk.

Regards,

Mike

http://www.mail-archive.com/time-nuts@febo.com/msg22283.html

Fri, 20 Nov 2009 John Larkin, of Highland Technology, was one of the most vocal advocates of not cutting the ground plane, as referenced in the first sentence at the top of this page.

However, he now cuts the ground plane in his latest products, then bridges the gap to the noisy plane with a few thin traces. However, since skin effect forces the high frequency noise currents to flow along the edges of the board, the currents on the noisy plane will still intermingle with the currents on the main ground plane.

This is not the same as leaving the original ground plane intact and having a separate plane that is attached only at one or only a few points. A separated, isolated ground plane, as described above, is far more effective in keeping the noise currents separated, and has none of the disadvantages of cutting the main ground plane.

Here's John's post to sci.electronics.design on Nov 20, 2009:

  From: John Larkin 
  Newsgroups: sci.electronics.design
  Subject: Re: day shot to hell
  Date: Fri, 20 Nov 2009 20:07:11 -0800
  Message-ID: 
  References: 
  
  <7mottlF3ifkudU1@mid.individual.net>
  Xref: Hurricane-Charley sci.electronics.design:656171
  X-Received-Date: Sat, 21 Nov 2009 04:07:12 UTC (s02-b42)

  On Fri, 20 Nov 2009 17:26:06 -0800,  Joerg 
  wrote:

  > don wrote:

  >> John Larkin wrote:

  >>> ftp://jjlarkin.lmi.net/First.JPG
  >>> ftp://jjlarkin.lmi.net/Swapped.jpg

  > Hey John, you wasted lots of real estate there. Wish I  could have
  > some of that. Doing an EMC fix on a client design right now  and I
  > can't even shove one more 0402 part in there :-(>

  We have  lots   of   space   for   this   one.   We're  replacing an
  older-generation board that is about 6x or so of our board area.

  You will  be not-pleased to know that the switcher  section  has its
  own rectangular  ground plane section that is connected to  the rest
  of the  plane  through  a number of thinnish  slivers.  The  rows of
  inductors, incoming and outgoing, straddle the plane gaps.  The idea
  is to  keep the various circulating currents in  the  switchers from
  leaking into the main ground plane where the analog stuff is.  I did
  the spread-spectrum thing on all the switchers, too.

  John

Here's the link to his post in google groups: http://groups.google.com/group/sci.electronics.design/msg/eb6c56847096e5f8