Capasitance of a comb track pcb does not add up! [Archive]

Wayne

May17-04, 07:56 PM

<jabberwocky><div class="vbmenu_control"><a href="jabberwocky:;" onClick="newWindow=window.open('','usenetCode','toolbar=no, location=no,scrollbars=yes,resizable=yes,status=no ,width=650,height=400'); newWindow.document.write('<HTML><HEAD><TITLE>Usenet ASCII</TITLE></HEAD><BODY topmargin=0 leftmargin=0 BGCOLOR=#F1F1F1><table border=0 width=625><td bgcolor=midnightblue><font color=#F1F1F1>This Usenet message\'s original ASCII form: </font></td></tr><tr><td width=449><br><br><font face=courier><UL><PRE>\n\nHi All\nI have built myself a printed circuit board that had interleaved\ntracks to produce a capacitor. (if you can not see the image I have\nenclosed then follow this link\nhttp://www.practicalcomponents.com/drawing_pages/pcb-saber.htm\nabout half way down you will see something similar.\n\n[Moderator\'s note: You won\'t see the image, because I removed it.\nBinary files are not welcome on sci.physics.research (or on other\nUsenet newsgroups, except those specifically designated for\nthem). -TB]\n\nThe characteristics are\n\nCopper track thickness 35um\ntrack length (fingers) 25mm (this is the total length facing the opposite\nelectrode)\ntrack width 0.45mm\nAnd there are 32 finger (31 interfaces)\n\nThis is on a material called FR4 that has a dielectric constant of 4-5. And\nair is apx 1.\n\nOK.\n\nThe copper is sitting on the track. So each finger of 25mm x 35u interface\nis separated by air.\n\nTherefore C_air=1.682E-11F\n\nHowever, if I measure the cap. I get 28pF (not 16.82pF).\n\nSo if I calculate it using FR4 dielectric constant I get 6.727E-11F.\nThis 67.27pF is much greater that measured in air and my empirical result.\n\nMy question is, as the copper is separated by air but sitting on FR4\n(dielectric const of 4) does this affect the overall capacitance. I am\ntrying to visualise it as a circular bands of magnetic waves - looking at it\nthis way it does seem that is must have an effect.\n\nRegards\n\nWayne Lawson\n\n\n</UL></PRE></font></td></tr></table></BODY><HTML>');"> <IMG SRC=/images/buttons/ip.gif BORDER=0 ALIGN=CENTER ALT="View this Usenet post in original ASCII form">  View this Usenet post in original ASCII form </a></div><P></jabberwocky>Hi All
I have built myself a printed circuit board that had interleaved
tracks to produce a capacitor. (if you can not see the image I have
enclosed then follow this link
http://www.practicalcomponents.com/drawing_pages/pcb-saber.htm
about half way down you will see something similar.

[Moderator's note: You won't see the image, because I removed it.
Binary files are not welcome on sci.physics.research (or on other
Usenet newsgroups, except those specifically designated for
them). -TB]

The characteristics are

Copper track thickness 35um
track length (fingers) 25mm (this is the total length facing the opposite
electrode)
track width .45mm
And there are 32 finger (31 interfaces)

This is on a material called FR4 that has a dielectric constant of 4-5. And
air is apx 1.

OK.

The copper is sitting on the track. So each finger of 25mm x 35u interface
is separated by air.

Therefore C_{air}=1.682E-11F

However, if I measure the cap. I get 28pF (not 16.82pF).

So if I calculate it using FR4 dielectric constant I get 6.727E-11F.
This 67.27pF is much greater that measured in air and my empirical result.

My question is, as the copper is separated by air but sitting on FR4
(dielectric const of 4) does this affect the overall capacitance. I am
trying to visualise it as a circular bands of magnetic waves - looking at it
this way it does seem that is must have an effect.

Regards

Wayne Lawson

Norm Dresner

May20-04, 11:47 AM

<jabberwocky><div class="vbmenu_control"><a href="jabberwocky:;" onClick="newWindow=window.open('','usenetCode','toolbar=no, location=no,scrollbars=yes,resizable=yes,status=no ,width=650,height=400'); newWindow.document.write('<HTML><HEAD><TITLE>Usenet ASCII</TITLE></HEAD><BODY topmargin=0 leftmargin=0 BGCOLOR=#F1F1F1><table border=0 width=625><td bgcolor=midnightblue><font color=#F1F1F1>This Usenet message\'s original ASCII form: </font></td></tr><tr><td width=449><br><br><font face=courier><UL><PRE>"Wayne" <nospam-mail@wlawson.com> wrote in message\nnews:haopc.41\\$XY4.1@newsfe1-win...\n>\n> Hi All\n> I have built myself a printed circuit board that had\ninterleaved\n> tracks to produce a capacitor. (if you can not see the image I have\n> enclosed then follow this link\n>\n> The characteristics are\n>\n> Copper track thickness 35um\n> track length (fingers) 25mm (this is the total length facing the opposite\n> electrode)\n> track width 0.45mm\n> And there are 32 finger (31 interfaces)\n>\n\nPerhaps if you shared the calculation method with us there might be some\nsuggestions made for the source of the inaccuracy.\n\nNorm\n\n</UL></PRE></font></td></tr></table></BODY><HTML>');"> <IMG SRC=/images/buttons/ip.gif BORDER=0 ALIGN=CENTER ALT="View this Usenet post in original ASCII form">  View this Usenet post in original ASCII form </a></div><P></jabberwocky>"Wayne" <nospam-mail@wlawson.com> wrote in message
news:haopc.41$XY4.1@newsfe1-win...
>
> Hi All
> I have built myself a printed circuit board that had
interleaved
> tracks to produce a capacitor. (if you can not see the image I have
> enclosed then follow this link
>
> The characteristics are
>
> Copper track thickness 35um
> track length (fingers) 25mm (this is the total length facing the opposite
> electrode)
> track width .45mm
> And there are 32 finger (31 interfaces)
>

Perhaps if you shared the calculation method with us there might be some
suggestions made for the source of the inaccuracy.

Norm

Patrick Van Esch

May20-04, 03:47 PM

<jabberwocky><div class="vbmenu_control"><a href="jabberwocky:;" onClick="newWindow=window.open('','usenetCode','toolbar=no, location=no,scrollbars=yes,resizable=yes,status=no ,width=650,height=400'); newWindow.document.write('<HTML><HEAD><TITLE>Usenet ASCII</TITLE></HEAD><BODY topmargin=0 leftmargin=0 BGCOLOR=#F1F1F1><table border=0 width=625><td bgcolor=midnightblue><font color=#F1F1F1>This Usenet message\'s original ASCII form: </font></td></tr><tr><td width=449><br><br><font face=courier><UL><PRE>"Wayne" <nospam-mail@wlawson.com> wrote in message news:<haopc.41\\$XY4.1@newsfe1-win>...\n\n>\n> The copper is sitting on the track. So each finger of 25mm x 35u interface\n> is separated by air.\n>\n> Therefore C_air=1.682E-11F\n>\n> However, if I measure the cap. I get 28pF (not 16.82pF).\n\n\nThe situation is much more complicated than what you seem to think.\nThe formula I think you apply is for the "infinite parallel plate"\ncapacitor, where the distance between the plates is much smaller\nthan the sides of the plates. But clearly here this is not the case.\nSo you get a rather complicated set of E-field lines, which induce\na charge pattern *on the surface* of your strips (and not only on the\nborder, as you seem to think). You\'d need a finite-element calculation\nto really find out what\'s the E-field configuration which can then\ngive you the capacity (namely the integrated charges over the strip\nsurfaces, which is nothing else but the E-field values at the surface\ntimes epsilon, divided by the line integral from one strip to another\nof the E-field (which is the same whatever path you choose) ).\n\nSo you shouldn\'t at all be surprised that you get wrong results when\napplying the epsilon S/d formula.\n\ncheers,\nPatrick.\n\n</UL></PRE></font></td></tr></table></BODY><HTML>');"> <IMG SRC=/images/buttons/ip.gif BORDER=0 ALIGN=CENTER ALT="View this Usenet post in original ASCII form">  View this Usenet post in original ASCII form </a></div><P></jabberwocky>"Wayne" <nospam-mail@wlawson.com> wrote in message news:<haopc.41$XY4.1@newsfe1-win>...

>
> The copper is sitting on the track. So each finger of 25mm x 35u interface
> is separated by air.
>
> Therefore C_{air}=1.682E-11F
>
> However, if I measure the cap. I get 28pF (not 16.82pF).

The situation is much more complicated than what you seem to think.
The formula I think you apply is for the "infinite parallel plate"
capacitor, where the distance between the plates is much smaller
than the sides of the plates. But clearly here this is not the case.
So you get a rather complicated set of E-field lines, which induce
a charge pattern *on the surface* of your strips (and not only on the
border, as you seem to think). You'd need a finite-element calculation
to really find out what's the E-field configuration which can then
give you the capacity (namely the integrated charges over the strip
surfaces, which is nothing else but the E-field values at the surface
times \epsilon, divided by the line integral from one strip to another
of the E-field (which is the same whatever path you choose) ).

So you shouldn't at all be surprised that you get wrong results when
applying the \epsilon S/d formula.

cheers,
Patrick.

Greg Neill

May20-04, 03:47 PM

<jabberwocky><div class="vbmenu_control"><a href="jabberwocky:;" onClick="newWindow=window.open('','usenetCode','toolbar=no, location=no,scrollbars=yes,resizable=yes,status=no ,width=650,height=400'); newWindow.document.write('<HTML><HEAD><TITLE>Usenet ASCII</TITLE></HEAD><BODY topmargin=0 leftmargin=0 BGCOLOR=#F1F1F1><table border=0 width=625><td bgcolor=midnightblue><font color=#F1F1F1>This Usenet message\'s original ASCII form: </font></td></tr><tr><td width=449><br><br><font face=courier><UL><PRE>"Norm Dresner" <ndrez@att.net> wrote in message\nnews:uzoqc.479\\$fF3.8111@bgtnsc05-news.ops.worldnet.att.net...\n> "Wayne" <nospam-mail@wlawson.com> wrote in message\n> news:haopc.41\\$XY4.1@newsfe1-win...\n> >\n> > Hi All\n> > I have built myself a printed circuit board that had\n> interleaved\n> > tracks to produce a capacitor. (if you can not see the image I have\n> > enclosed then follow this link\n> >\n> > The characteristics are\n> >\n> > Copper track thickness 35um\n> > track length (fingers) 25mm (this is the total length facing the\nopposite\n> > electrode)\n> > track width 0.45mm\n> > And there are 32 finger (31 interfaces)\n> >\n>\n> Perhaps if you shared the calculation method with us there might be some\n> suggestions made for the source of the inaccuracy.\n>\n\nAlso, try sandwiching another blank piece of PC board\non top of the capacitor tracks (touching it, if possible).\nSee if the capacitance goes up when the piece is in place.\n\n</UL></PRE></font></td></tr></table></BODY><HTML>');"> <IMG SRC=/images/buttons/ip.gif BORDER=0 ALIGN=CENTER ALT="View this Usenet post in original ASCII form">  View this Usenet post in original ASCII form </a></div><P></jabberwocky>"Norm Dresner" <ndrez@att.net> wrote in message
news:uzoqc.479$fF3.8111@bgtnsc05-news.ops.worldnet.att.net...
> "Wayne" <nospam-mail@wlawson.com> wrote in message
> news:haopc.41$XY4.1@newsfe1-win...
> >
> > Hi All
> > I have built myself a printed circuit board that had
> interleaved
> > tracks to produce a capacitor. (if you can not see the image I have
> > enclosed then follow this link
> >
> > The characteristics are
> >
> > Copper track thickness 35um
> > track length (fingers) 25mm (this is the total length facing the
opposite
> > electrode)
> > track width .45mm
> > And there are 32 finger (31 interfaces)
> >
>
> Perhaps if you shared the calculation method with us there might be some
> suggestions made for the source of the inaccuracy.
>

Also, try sandwiching another blank piece of PC board
on top of the capacitor tracks (touching it, if possible).
See if the capacitance goes up when the piece is in place.