PCBs with ceramic substrates
There's little out there on the frequency dependence of dielectric constant at audio and sub audio frequencies.
A bit more for 10MHz and above like:
http://www.edn.com/design/analog/4398951/2/What-PCB-material-do-I-need-to-use-for-RF-
http://ac.els-cdn.com/S1474667015373043/1-s2.0-S1474667015373043-main.pdf?_tid=2e1d6d30-231a-11e7-84ab-00000aab0f26&acdnat=1492398280_a0295abea04d2b08db9f227ee52506fc http://ac.els-cdn.com/S1474667015373043/1-s2.0-S1474667015373043-main.pdf?_tid=2e1d6d30-231a-11e7-84ab-00000aab0f26&acdnat=1492398280_a0295abea04d2b08db9f227ee52506fc
There's also the article in The October 8 1978 issue of electronics on
Getting rid of hook: The hidden pc-board capacitance by authors from Tektronix and Norplex.
This article is in the files section of the Tekscopes yahoo group.
Otherwise there are various values given in the literature and on the web for the low frequency relative permittivity of FR4 which are all noticeably higher than the value at 10MHz.
Bruce
On 17 April 2017 at 14:31 cheater00 cheater00 cheater00@gmail.com wrote:
Thank you. Do you have a plot of the effect on impedance due to hook out into high frequencies, measured on a real world material? It would tell me a lot.
On Mon, 17 Apr 2017 04:19 Bruce Griffiths, <bruce.griffiths@xtra.co.nz mailto:bruce.griffiths@xtra.co.nz > wrote:
Hook is merely a manifestation of the variation of dielectric constant with frequency.
It affects the frequency response and transient of resistive attenuators by requiring more complex compensation than merely adjusting a trimmer cap to equalise the low frequency and high frequency attenuation. Since the dielectric "constant" (relative permittivity) of all dielectrics is frequency dependent all dielectrics will exhibit hook to some extent. Some PCB substrates like some versions of FR4 and G10 exhibit a significant variation in the dielectric constant from the dc value to a somewhat lower value for frequencies even in he audio range let alone frequencies of several MHz. Achieving a flat frequency response where the dielectric associated with circuit board capacitances exhibits significant hook is a complex task. Circuit board hook even affects the impedance of printed transmission lines (eg stripline, microstrip, CPW etc).
Bruce
On 17 April 2017 at 13:54 cheater00 cheater00 <cheater00@gmail.com mailto:cheater00@gmail.com > wrote:
Are conformal coatings the right way to handle this?
I understand there are kinds of FR4 and G10 that don't have hook. What does
one do about hook - how are those substrates improved? How does hook
manifest in circuits?
On Mon, 17 Apr 2017 03:48 David, <davidwhess@gmail.com mailto:davidwhess@gmail.com > wrote:
FR4 has problems with consistency. Samples can have problems with
hook, dielectric absorption, leakage, and sensitivity to humidity.
On Sun, 16 Apr 2017 12:08:07 +0100, you wrote:
Hi Chuck
But the context is "PCBs with ceramic substrates". Are any of *those*
tough? They may well be, perhaps you know of some? It does not help us
with the subject much if there are ceramics with these amazing
properties if they are not available as PCBs.
There is also the question of exactly what properties of FR4 are
limiting for "metrology" use.
John
_______________________________________________
volt-nuts mailing list -- volt-nuts@febo.com mailto:volt-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.
_______________________________________________
volt-nuts mailing list -- volt-nuts@febo.com mailto:volt-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.
Thanks, that second article had exactly what I had in mind. I would love to
see that permittivity vs freq graph in more detail around one of the flat
crossings of the saw graph with the smooth fitted curve, maybe +/- 10 MHz
around that crossing (not around the steep jump)
If ceramic is fragile and FR4 can be hook less then I assume FR4 is better,
testing anyways doesn't hurt.
The periodicity of the "saw" graph is telling. Could this have to do with
the size of layers created by the fibers, or the size of fibers themselves
or their spacing? Maybe a random fiber dimension (diameter, spacing) would
be better here than using a constant dimension of fibers.
On Mon, 17 Apr 2017 05:35 Bruce Griffiths, bruce.griffiths@xtra.co.nz
wrote:
There's little out there on the frequency dependence of dielectric
constant at audio and sub audio frequencies.
A bit more for 10MHz and above like:
http://www.edn.com/design/analog/4398951/2/What-PCB-material-do-I-need-to-use-for-RF-
There's also the article in The October 8 1978 issue of electronics on
Getting rid of hook: The hidden pc-board capacitance by authors from
Tektronix and Norplex.
This article is in the files section of the Tekscopes yahoo group.
Otherwise there are various values given in the literature and on the web
for the low frequency relative permittivity of FR4 which are all noticeably
higher than the value at 10MHz.
Bruce
On 17 April 2017 at 14:31 cheater00 cheater00 cheater00@gmail.com wrote:
Thank you. Do you have a plot of the effect on impedance due to hook out
into high frequencies, measured on a real world material? It would tell me
a lot.
On Mon, 17 Apr 2017 04:19 Bruce Griffiths, bruce.griffiths@xtra.co.nz
wrote:
Hook is merely a manifestation of the variation of dielectric constant
with frequency.
It affects the frequency response and transient of resistive attenuators
by requiring more complex compensation than merely adjusting a trimmer cap
to equalise the low frequency and high frequency attenuation. Since the
dielectric "constant" (relative permittivity) of all dielectrics is
frequency dependent all dielectrics will exhibit hook to some extent. Some
PCB substrates like some versions of FR4 and G10 exhibit a significant
variation in the dielectric constant from the dc value to a somewhat lower
value for frequencies even in he audio range let alone frequencies of
several MHz. Achieving a flat frequency response where the dielectric
associated with circuit board capacitances exhibits significant hook is a
complex task. Circuit board hook even affects the impedance of printed
transmission lines (eg stripline, microstrip, CPW etc).
Bruce
On 17 April 2017 at 13:54 cheater00 cheater00 cheater00@gmail.com wrote:
Are conformal coatings the right way to handle this?
I understand there are kinds of FR4 and G10 that don't have hook. What does
one do about hook - how are those substrates improved? How does hook
manifest in circuits?
On Mon, 17 Apr 2017 03:48 David, davidwhess@gmail.com wrote:
FR4 has problems with consistency. Samples can have problems with
hook, dielectric absorption, leakage, and sensitivity to humidity.
On Sun, 16 Apr 2017 12:08:07 +0100, you wrote:
Hi Chuck
But the context is "PCBs with ceramic substrates". Are any of those
tough? They may well be, perhaps you know of some? It does not help us
with the subject much if there are ceramics with these amazing
properties if they are not available as PCBs.
There is also the question of exactly what properties of FR4 are
limiting for "metrology" use.
John
volt-nuts mailing list -- volt-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.
volt-nuts mailing list -- volt-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.
When the glass fibre bunch periodicity approaches a significant fraction of a wavelength the periodic spatial modulation of the substrate dielectric constant under or surrounding transmission lines has interesting effects on the properties of the transmission line. At even higher frequencies the surface texture of the copper also becomes an issue. Ditching glass fibre in favour of small ceramic particles eliminates the effect of the periodic spatial modulation of the substrate dielectric constant. Just as with dielectric absorption it should be possible to fit a Cole-Coles or similar model to the variation of dielectric properties (including loss) with frequency.
Bruce
On 17 April 2017 at 15:54 cheater00 cheater00 cheater00@gmail.com wrote:
Thanks, that second article had exactly what I had in mind. I would love to see that permittivity vs freq graph in more detail around one of the flat crossings of the saw graph with the smooth fitted curve, maybe +/- 10 MHz around that crossing (not around the steep jump)
If ceramic is fragile and FR4 can be hook less then I assume FR4 is better, testing anyways doesn't hurt.
The periodicity of the "saw" graph is telling. Could this have to do with the size of layers created by the fibers, or the size of fibers themselves or their spacing? Maybe a random fiber dimension (diameter, spacing) would be better here than using a constant dimension of fibers.
On Mon, 17 Apr 2017 05:35 Bruce Griffiths, <bruce.griffiths@xtra.co.nz mailto:bruce.griffiths@xtra.co.nz > wrote:
There's little out there on the frequency dependence of dielectric constant at audio and sub audio frequencies.
A bit more for 10MHz and above like:
http://www.edn.com/design/analog/4398951/2/What-PCB-material-do-I-need-to-use-for-RF-
http://ac.els-cdn.com/S1474667015373043/1-s2.0-S1474667015373043-main.pdf?_tid=2e1d6d30-231a-11e7-84ab-00000aab0f26&acdnat=1492398280_a0295abea04d2b08db9f227ee52506fc http://ac.els-cdn.com/S1474667015373043/1-s2.0-S1474667015373043-main.pdf?_tid=2e1d6d30-231a-11e7-84ab-00000aab0f26&acdnat=1492398280_a0295abea04d2b08db9f227ee52506fc
There's also the article in The October 8 1978 issue of electronics on
Getting rid of hook: The hidden pc-board capacitance by authors from Tektronix and Norplex.
This article is in the files section of the Tekscopes yahoo group.
Otherwise there are various values given in the literature and on the web for the low frequency relative permittivity of FR4 which are all noticeably higher than the value at 10MHz.
Bruce
On 17 April 2017 at 14:31 cheater00 cheater00 <cheater00@gmail.com mailto:cheater00@gmail.com > wrote:
Thank you. Do you have a plot of the effect on impedance due to hook out into high frequencies, measured on a real world material? It would tell me a lot.
On Mon, 17 Apr 2017 04:19 Bruce Griffiths, <bruce.griffiths@xtra.co.nz mailto:bruce.griffiths@xtra.co.nz > wrote:
Hook is merely a manifestation of the variation of dielectric constant with frequency.
It affects the frequency response and transient of resistive attenuators by requiring more complex compensation than merely adjusting a trimmer cap to equalise the low frequency and high frequency attenuation. Since the dielectric "constant" (relative permittivity) of all dielectrics is frequency dependent all dielectrics will exhibit hook to some extent. Some PCB substrates like some versions of FR4 and G10 exhibit a significant variation in the dielectric constant from the dc value to a somewhat lower value for frequencies even in he audio range let alone frequencies of several MHz. Achieving a flat frequency response where the dielectric associated with circuit board capacitances exhibits significant hook is a complex task. Circuit board hook even affects the impedance of printed transmission lines (eg stripline, microstrip, CPW etc).
Bruce
On 17 April 2017 at 13:54 cheater00 cheater00 <cheater00@gmail.com mailto:cheater00@gmail.com > wrote:
Are conformal coatings the right way to handle this?
I understand there are kinds of FR4 and G10 that don't have hook. What does
one do about hook - how are those substrates improved? How does hook
manifest in circuits?
On Mon, 17 Apr 2017 03:48 David, <davidwhess@gmail.com mailto:davidwhess@gmail.com > wrote:
FR4 has problems with consistency. Samples can have problems with
hook, dielectric absorption, leakage, and sensitivity to humidity.
On Sun, 16 Apr 2017 12:08:07 +0100, you wrote:
Hi Chuck
But the context is "PCBs with ceramic substrates". Are any of *those*
tough? They may well be, perhaps you know of some? It does not help us
with the subject much if there are ceramics with these amazing
properties if they are not available as PCBs.
There is also the question of exactly what properties of FR4 are
limiting for "metrology" use.
John
_______________________________________________
volt-nuts mailing list -- volt-nuts@febo.com mailto:volt-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.
_______________________________________________
volt-nuts mailing list -- volt-nuts@febo.com mailto:volt-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.
Indeed, someone I asked elsewhere hinted that ceramic filled PTFE is a good
idea for curbing hook.
On Mon, 17 Apr 2017 06:20 Bruce Griffiths, bruce.griffiths@xtra.co.nz
wrote:
When the glass fibre bunch periodicity approaches a significant fraction
of a wavelength the periodic spatial modulation of the substrate dielectric
constant under or surrounding transmission lines has interesting effects
on the properties of the transmission line. At even higher frequencies the
surface texture of the copper also becomes an issue. Ditching glass fibre
in favour of small ceramic particles eliminates the effect of the periodic
spatial modulation of the substrate dielectric constant. Just as with
dielectric absorption it should be possible to fit a Cole-Coles or similar
model to the variation of dielectric properties (including loss) with
frequency.
Bruce
On 17 April 2017 at 15:54 cheater00 cheater00 cheater00@gmail.com wrote:
Thanks, that second article had exactly what I had in mind. I would love
to see that permittivity vs freq graph in more detail around one of the
flat crossings of the saw graph with the smooth fitted curve, maybe +/- 10
MHz around that crossing (not around the steep jump)
If ceramic is fragile and FR4 can be hook less then I assume FR4 is
better, testing anyways doesn't hurt.
The periodicity of the "saw" graph is telling. Could this have to do with
the size of layers created by the fibers, or the size of fibers themselves
or their spacing? Maybe a random fiber dimension (diameter, spacing) would
be better here than using a constant dimension of fibers.
On Mon, 17 Apr 2017 05:35 Bruce Griffiths, bruce.griffiths@xtra.co.nz
wrote:
There's little out there on the frequency dependence of dielectric
constant at audio and sub audio frequencies.
A bit more for 10MHz and above like:
http://www.edn.com/design/analog/4398951/2/What-PCB-material-do-I-need-to-use-for-RF-
There's also the article in The October 8 1978 issue of electronics on
Getting rid of hook: The hidden pc-board capacitance by authors from
Tektronix and Norplex.
This article is in the files section of the Tekscopes yahoo group.
Otherwise there are various values given in the literature and on the web
for the low frequency relative permittivity of FR4 which are all noticeably
higher than the value at 10MHz.
Bruce
On 17 April 2017 at 14:31 cheater00 cheater00 cheater00@gmail.com wrote:
Thank you. Do you have a plot of the effect on impedance due to hook out
into high frequencies, measured on a real world material? It would tell me
a lot.
On Mon, 17 Apr 2017 04:19 Bruce Griffiths, bruce.griffiths@xtra.co.nz
wrote:
Hook is merely a manifestation of the variation of dielectric constant
with frequency.
It affects the frequency response and transient of resistive attenuators
by requiring more complex compensation than merely adjusting a trimmer cap
to equalise the low frequency and high frequency attenuation. Since the
dielectric "constant" (relative permittivity) of all dielectrics is
frequency dependent all dielectrics will exhibit hook to some extent. Some
PCB substrates like some versions of FR4 and G10 exhibit a significant
variation in the dielectric constant from the dc value to a somewhat lower
value for frequencies even in he audio range let alone frequencies of
several MHz. Achieving a flat frequency response where the dielectric
associated with circuit board capacitances exhibits significant hook is a
complex task. Circuit board hook even affects the impedance of printed
transmission lines (eg stripline, microstrip, CPW etc).
Bruce
On 17 April 2017 at 13:54 cheater00 cheater00 cheater00@gmail.com wrote:
Are conformal coatings the right way to handle this?
I understand there are kinds of FR4 and G10 that don't have hook. What does
one do about hook - how are those substrates improved? How does hook
manifest in circuits?
On Mon, 17 Apr 2017 03:48 David, davidwhess@gmail.com wrote:
FR4 has problems with consistency. Samples can have problems with
hook, dielectric absorption, leakage, and sensitivity to humidity.
On Sun, 16 Apr 2017 12:08:07 +0100, you wrote:
Hi Chuck
But the context is "PCBs with ceramic substrates". Are any of those
tough? They may well be, perhaps you know of some? It does not help us
with the subject much if there are ceramics with these amazing
properties if they are not available as PCBs.
There is also the question of exactly what properties of FR4 are
limiting for "metrology" use.
John
volt-nuts mailing list -- volt-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.
volt-nuts mailing list -- volt-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.