[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Grounding schemes was:A prospective s.s.t.c. input-synchronizer (fwd)



Original poster: Gerry Reynolds <greynolds@xxxxxxxxxx>



---------- Forwarded message ----------
Date: Thu, 21 Dec 2006 16:52:14 -0700
From: Gerry  Reynolds <gerryreynolds@xxxxxxxxxxxxx>
To: Tesla list <tesla@xxxxxxxxxx>
Subject: Re: Grounding schemes was:A prospective s.s.t.c. input-synchronizer
    (fwd)

Hi Steve,

Oh GOOD,  grounding issues can be such a confusing issue even for a lot of 
engineers.  Often problems can occur when the return current for a signal is 
not allowed to follow the signal path (such as a lack of a ground plane or 
going thru a connector, etc) and when mixing high current circuits with low 
current circuits (such as not star grounding, etc).  If one understands 
where the currents go,  the IR drops the high currents can create in the 
conductor and how this can affect other grounds, the signal current paths 
and return current paths, a lot is gained in understanding and heading off 
potential problems.

Take care,
Gerry R

> ---------- Forwarded message ----------
> Date: Thu, 21 Dec 2006 11:26:35 -0600
> From: Steve Ward <steve.ward@xxxxxxxxx>
> To: Tesla list <tesla@xxxxxxxxxx>
> Subject: Re: Grounding schemes was:A prospective s.s.t.c. 
> input-synchronizer
>    (fwd)
>
> Hi Gerry,
>
> My comment about the star grounding was mentioned not as the actual
> technique for each of the ground points for the 5V logic chips.  On
> the actual board i too use a planar ground scheme.  But, this ground
> plane is then tied to a single point, where other parts of my DRSSTC
> are grounded.  What i wanted to specify was that some higher current
> circuit didnt have its ground return flowing through the 5V ground
> plane, but rather, all circuits requireing ground simply tied to a
> point, which is then connected to mains ground.
>
> I hope that is clear... i have trouble explaining it properly without 
> pictures.
>
> Steve Ward
>
> On 12/20/06, Tesla list <tesla@xxxxxxxxxx> wrote:
>> Original poster: Gerry Reynolds <greynolds@xxxxxxxxxx>
>>
>>
>>
>> ---------- Forwarded message ----------
>> Date: Wed, 20 Dec 2006 00:30:16 -0700
>> From: Gerry  Reynolds <gerryreynolds@xxxxxxxxxxxxx>
>> To: Tesla list <tesla@xxxxxxxxxx>
>> Subject: Re: Grounding schemes  was:A prospective s.s.t.c. 
>> input-synchronizer
>>     (fwd)
>>
>> Hi Steve,
>>
>> Star grounding vs a ground plane:
>>
>> Digital logic signals routed over a ground plane will result in return
>> currents that follow the path of the signal route.  Any drop in the 
>> ground
>> plane caused by the return current will be specific to the path the 
>> return
>> current follows (for the most part assuming the conductivity of the 
>> ground
>> plane is reasonably good).  For logic,  ground noise is not important up 
>> to
>> a point where thresholds of switching are violated.
>>
>> For analog signals, ground noise can be interpreted as a signal component
>> and amplified accordingly and star grounding can help control noise
>> coupling.  However, it will be important that the signal path follows the
>> ground path so the return current can follow the signal path.  Otherwise, 
>> a
>> "loop" is formed making the signal/gound pair subceptible to outside
>> influences (like magnetic fields).  It has been my experience that ground
>> planes give the best grounding experience with most situations.
>>
>> Gerry R.
>>
>>
>> > Hi Ken,
>> >
>> > It may be either a grounding issue, or it could be possible that the
>> > noise isnt real.  Putting ceramic caps on every IC V+ and ground
>> > should normally be good enough.  Make sure that you dont have any high
>> > currents flowing through the 5V ground.  If the 5V circuit is grounded
>> > to anything else, be sure to use "star" grounding, and the 5V circuit
>> > is just a branch on the star, and not the center.
>> >
>> > You probably already know all this, but you did welcome suggestions,
>> > and hopefully it will get you started on figuring out the problem.
>> >
>> > Steve
>> >
>> > On 12/19/06, Tesla list <tesla@xxxxxxxxxx> wrote:
>> >> Original poster: Gerry Reynolds <greynolds@xxxxxxxxxx>
>> >>
>> >>
>> >>
>> >> ---------- Forwarded message ----------
>> >> Date: Tue, 19 Dec 2006 11:28:54 -0800
>> >> From: K. C. Herrick <kchdlh@xxxxxxx>
>> >> To: Tesla List <tesla@xxxxxxxxxx>
>> >> Subject: Re: A prospective s.s.t.c. input-synchronizer
>> >>
>> >> Chip, Steve Ward (& all)-
>> >> ----------------------------------------------------------
>> >>
>> >> Hi Ken.
>> >>
>> >> Sorry for the delay.  I got slammed at work and didn't end up on the
>> >> computer last night.  Here's the link.  It's just plain right now but 
>> >> can
>> >> be fixed up to have html in it later. I just wanted to get it done.
>> >> Sorry
>> >> for the delay again.
>> >>
>> >> Chip
>> >>
>> >> http://www.pupman.com/current/kcherrick/tch-synctest4a.jpg
>> >> ------------------------------------------
>> >>
>> >> No problemo; looks perfect as-is.  And to all:  The above link is the
>> >> replacement circuit for my synctest2.jpg that I'd posted
>> >> previously on hot-streamer.  This one seems to work in the hardware.
>> >>
>> >>
>> >> And Steve...in further response to yours of 12/18: My low-level EMI
>> >> problem appears to stem from a 14 MHz transient that appears on the 
>> >> +5V
>> >> bus every time the H-bridge switches.  I'm now trying to track that
>> >> down.  Suggestions on that from >>anyone<< are welcome!!
>> >>
>> >> KCH
>> >>
>> >>
>> >>
>> >>
>> >
>> >
>> >
>> >
>>
>>
>>
>>
>>
>
>
>
>