| Author |
Message |
gmv
Guest
|
 Posted:
Sat Feb 05, 2005 2:33 am Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
Sorry, Do not know what a DSP/FIR is.
I am using only 8 bit data.
Surprising enough the filters work quite well
between 1Hz and DC just like they do between
1Hz and infinity. The only problem I have is
getting those capacitors in a non-electrolytic form.
I have found I can get pretty good matches in frequency
simply if I can know the true value of the capacitors.
The design matches the theory if only I can
match the capacitors.
I used to laugh at precision and accuracy but
I have found that is the only way to go when
dealing with filter circuits and possibly
oscillator circuits too.
The only way I will play with things like FIR or
DSP is if it will handle data from a non-linear
kind of sensor that tends to overdrive the
preamp at high freqs making it impossible to
look at the low freqs also it got to be cheap.
I need to filter out
the upper freqs right off the bat or it is
impossible to get enough gain to look at the low.
Yes, a log amp is a possibility, I have already
considered that. But the signals I now receive
are good enough for myself and all I wish to do
is to bring closure to my circuit be acquiring what
I consider to be decent capacitors. These electrolytics
are seemingly only for one frequency because
funny things start happening if you read the
capacitance with different frequencies.
Like I get a lower capacitance at a higher frequency.
I have found no such problems with regular
caps that do not use an electrolytic form.
Your solution would be possible if I had a sensor
resonant at 20 seconds but the Geophone I am using
is resonant at something over 1 Hz.
Geospace has just the device I would like to have
they have a .5 Hz Geophone but they want about
$3000 for one and that is way too high for someone
living on Social Security.
"Rob Gaddi" <rgaddi@bcm.YUMMYSPAMtmc.edu> wrote in message news:cu0elc$t4c@gazette.corp.bcm.tmc.edu...
| Quote: | While I hate to give my usual answer, have you considered just digitizing the data and running it through a DSP? You're talking
about some massively slow filters to be trying to make with any reliable tolerance in analog, and with component values that large
it's not going to be cheap. Plus you can use a linear FIR filter and not have to worry about the phase distortion you were
mentioning.
If your noise is really swamping the signal, and dynamic range is a concern, at the low speeds you're talking about you could run
the signal through a log amp before you A/D it.
Lucky for me different phases of Earthquakes
seem to prefer different frequencies.
But my picture will never be as good as the
big boys.
Just too darn noisy at this location.
I get a terrible noise at maybe 15 Hz or so.
Need to filter like crazy to get rid of it.
It drowns out just about everything i try to do.
I suspect but can not prove that my geophone
is out of spec and not working at the 1Hz
it is supposed to.
The only adjustment I can make is to center the
mass in the housing. If there is a way to lower the
resonant freq I would sure like to know.
I am thinking of finding some weights to put on
the adjustment screw as much weight as can still be
centered.
But it is my experience that for a spring to
lower its frequency it needs to extend or compress
farther in distance. similar to a pendulums length.
T=2PiSqRt(length/gravity) |
|
|
| Back to top |
|
 |
gmv
Guest
|
| Posted:
Sat Feb 05, 2005 2:39 am Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
"R.Lewis" <h.lewis@connect-2.co.uk> wrote in message news:36htfrF5289nrU1@individual.net...
| Quote: |
"gmv" <noemail@please.123> wrote in message
news:HgIMd.47990$EG1.27201@attbi_s53...
"Watson A.Name - "Watt Sun, the Dark Remover"" <NOSPAM@dslextreme.com
wrote in message news:1106d0pdpsb0i35@corp.supernews.com...
"gmv" <noemail@please.123> wrote in message
news:PczMd.46765$IV5.32350@attbi_s54...
Hello Rob,
I have sent an email to the Yuden people they seem to
have exactly what i need but it is not in a leaded
package I have hopes I can make a leaded cap by soldering
wires on each end of the surface mount type.
The real trouble here is I need the caps to be
closely matched so I am afraid I will have to
buy quite a few just to get a couple to match.
Matching is more important then having the exact value here
because I am making bandpass filters that require
fairly close tolerances.
The usual technique is to raise the impedance of the circuit
so that the capacitors can be lower values, below 1 uF. I
realize inductors can get very large and unwieldy, but if you
use RC filters such as twin-T or Wien bridge, the capacitors
can be reasonably sized, below 1 uF.
I would like to keep the resistors as small as I can to
lower the noise they introduce into the system
I am using quite high gain maybe 100 to 105dbv
so that is why I am using large capacitors.
I am amplifying a Geophone signal then feeding
the signal to three different narrow band filters
then summing those three filters into a low pass
filter amplifier for the final run. I am getting
interesting results even though the spice program
indicates terrible phase distortion.
I am fighting terrible noise at my location and
this seems to be the only way to see anything.
Lucky for me different phases of Earthquakes
seem to prefer different frequencies.
But my picture will never be as good as the
big boys.
Just too darn noisy at this location.
I get a terrible noise at maybe 15 Hz or so.
Need to filter like crazy to get rid of it.
It drowns out just about everything i try to do.
I suspect but can not prove that my geophone
is out of spec and not working at the 1Hz
it is supposed to.
The only adjustment I can make is to center the
mass in the housing. If there is a way to lower the
resonant freq I would sure like to know.
I am thinking of finding some weights to put on
the adjustment screw as much weight as can still be
centered.
But it is my experience that for a spring to
lower its frequency it needs to extend or compress
farther in distance. similar to a pendulums length.
T=2PiSqRt(length/gravity)
Another problem is I need only about 30 caps
but they sell them by the zillions.
Why in heck do the US people not carry
non-electrolytic caps in these ranges
of 1uf to 100uf ?
Thanks for your lead on this company.
--
Regards;
gmv
"Rob Gaddi" <rgaddi@bcm.YUMMYSPAMtmc.edu> wrote in message
news:ctu9op$nud@gazette.corp.bcm.tmc.edu...
snip
Run it through an A/D to get it into the digital domain where you can handle
it sensibly without having to consider what temperature, sonics, etc will
otherwise do to your filters if you need that sort of accuracy.
|
I must filter out the high stuff first because my sample rate
is only 65539.3 samples per hour and my dynamic range is only
8 bits. Like I told someone else my sensor has a non-linear
response in the frequency range of interest.
2Hz to 1/20 Hz.
It would be so wonderful to have a dead end at 2HZ kind
of filter then most of my troubles would evaporate. |
|
| Back to top |
|
|
gmv
Guest
|
| Posted:
Sat Feb 05, 2005 2:47 am Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
They are really pretty good the tolerance refers
to the actual value of the cap verses the
indicated value but there seems to be funny
idiosyncrasies related to using the electrolytic
stuff, so to speak, like the capacitance seems to change
with the frequency....maybe I just have a bad meter
to conduct measurements with but I am certain it tells
me the truth for non-electrolytic types of capacitors.
It seems if you want to work near 1Hz you should measure
the capacitance with a 1Hz meter but I know of no such animal.
"Jim Backus" <jhb@nospam.co.uk> wrote in message news:TpquPuPd0tCd-pn2-SgulLeF9nLcd@localhost...
| Quote: | On Thu, 3 Feb 2005 21:37:45 UTC, "gmv" <noemail@please.123> wrote:
I am trying to get a reading
within 1% of reality.
I'm not sure reality allows measuring electrolytics to 1% :-)
I've never considered electrolytics to be stable components. Their
typical manufacturing tolerance is -20% +50%.
--
Jim Backus OS/2 user since 1994
bona fide replies to j <dot> backus <the circle thingy> jita <dot
demon <dot> co <dot> uk
|
|
|
| Back to top |
|
|
Watson A.Name - \"Watt Su
Guest
|
| Posted:
Sat Feb 05, 2005 3:40 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
"gmv" <noemail@please.123> wrote in message
news:CAQMd.48480$EG1.21315@attbi_s53...
| Quote: | I have a fear in dealing with oscillators
I have no ham license and do not want to
generate more RF energy then currently exists.
I choose to go passive and look only at
baseband signals.
If I wanted to do things right i would
use a capacitive sensor in a kind of
bridge configuration and the capacitor would
FM or AM a carrier frequency
to be later demodulated and amplified.
But like I said there already is too much
RFI and I just do not want to deal with the FCC.
This country might use such an excuse to
throw me back in the nuthouse for the fifth time.
|
Last summer I built a proximity detector that ran at about 450 to 500
kHz. I was disappointed at its sensitivity, only worked to about 3/4"(2
cm). It radiated some RF, but at that freq. the AM radio could only
pick it up at arm's length. This sensor used a capacitive plate to damp
out the oscillator, but the circuit could be changed to vary the freq as
the capacitance changed. In other words, it could have made a good
position sensor.
Anyway, what you need to do is use a mechanical filter, instead of
trying to make an electronic filter. One thought comes to mind. Use a
speaker and microphone coupled by a mechanical link, and add weight to
that link so that it would only respond to freqs below some low value,
such as 1 Hz. This also makes it a bandpass filter.
| Quote: | "Watson A.Name - "Watt Sun, the Dark Remover"" <NOSPAM@dslextreme.com
wrote in message news:1106mgko9mlpiac@corp.supernews.com...
"gmv" <noemail@please.123> wrote in message
news:%nIMd.48003$EG1.44889@attbi_s53...
They dont work at freqs below 10Hz
unless they are very large.
I have never seen a 1/20Hz crystal before.
You're right, a xtal would have to be huge to get down to that freq.
However you could multiply the low freq many times and then filter
it.
"loedown" <zero242@internode.on.net> wrote in message
news:42034d8b$1@duster.adelaide.on.net...
If you need really closely matched bandpass filters, have you
considered crystal filters. I know very little about them, but I do
know they have a very sharp cutoff / roll off
Paul
|
|
|
| Back to top |
|
|
Watson A.Name - \"Watt Su
Guest
|
| Posted:
Sat Feb 05, 2005 3:58 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
"Jim Backus" <jhb@nospam.co.uk> wrote in message
news:TpquPuPd0tCd-pn2-SgulLeF9nLcd@localhost...
| Quote: | On Thu, 3 Feb 2005 21:37:45 UTC, "gmv" <noemail@please.123> wrote:
I am trying to get a reading
within 1% of reality.
I'm not sure reality allows measuring electrolytics to 1% :-)
I've never considered electrolytics to be stable components. Their
typical manufacturing tolerance is -20% +50%.
--
Jim Backus OS/2 user since 1994
|
But those ceramic caps that are .1 uF or more have tempcos that seem to
make them great temp sensors. I put my fingers on a cold one and it
changes by ten percent or more. Makes me think that I could use one in
a chirping cricket circuit and just like a cricket, it will speed up as
it gets warmer. |
|
| Back to top |
|
|
Watson A.Name - \"Watt Su
Guest
|
| Posted:
Sat Feb 05, 2005 4:25 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
"gmv" <noemail@please.123> wrote in message
news:f_QMd.49525$eT5.27123@attbi_s51...
| Quote: | Sorry, Do not know what a DSP/FIR is.
I am using only 8 bit data.
Surprising enough the filters work quite well
between 1Hz and DC just like they do between
1Hz and infinity. The only problem I have is
getting those capacitors in a non-electrolytic form.
I have found I can get pretty good matches in frequency
simply if I can know the true value of the capacitors.
The design matches the theory if only I can
match the capacitors.
I used to laugh at precision and accuracy but
I have found that is the only way to go when
dealing with filter circuits and possibly
oscillator circuits too.
The only way I will play with things like FIR or
DSP is if it will handle data from a non-linearI still think
kind of sensor that tends to overdrive the
preamp at high freqs making it impossible to
look at the low freqs also it got to be cheap.
I need to filter out
the upper freqs right off the bat or it is
impossible to get enough gain to look at the low.
|
I still think you need to raise the impedance of the circuits
up high enough so that the capacitors are much smaller.
Using FET input opamps and megohm resistors, you should be able
to get down to very low freqs with caps of a microfarad.
I offered that Action Electronics web page with the motor capacitors,
but I don't know if they are electrolytic or not. Maybe someone
could help. They also have tantalom caps, which are more stable
than 'lytics. http://www.action-electronics.com/capacitr.htm
| Quote: | Yes, a log amp is a possibility, I have already
considered that. But the signals I now receive
are good enough for myself and all I wish to do
is to bring closure to my circuit be acquiring what
I consider to be decent capacitors. These electrolytics
are seemingly only for one frequency because
funny things start happening if you read the
capacitance with different frequencies.
Like I get a lower capacitance at a higher frequency.
I have found no such problems with regular
caps that do not use an electrolytic form.
Your solution would be possible if I had a sensor
resonant at 20 seconds but the Geophone I am using
is resonant at something over 1 Hz.
Geospace has just the device I would like to have
they have a .5 Hz Geophone but they want about
$3000 for one and that is way too high for someone
living on Social Security.
"Rob Gaddi" <rgaddi@bcm.YUMMYSPAMtmc.edu> wrote in message
news:cu0elc$t4c@gazette.corp.bcm.tmc.edu...
While I hate to give my usual answer, have you considered just
digitizing the data and running it through a DSP? You're talking
about some massively slow filters to be trying to make with any
reliable tolerance in analog, and with component values that large
it's not going to be cheap. Plus you can use a linear FIR filter
and not have to worry about the phase distortion you were
mentioning.
If your noise is really swamping the signal, and dynamic range is a
concern, at the low speeds you're talking about you could run
the signal through a log amp before you A/D it.
Lucky for me different phases of Earthquakes
seem to prefer different frequencies.
But my picture will never be as good as the
big boys.
Just too darn noisy at this location.
I get a terrible noise at maybe 15 Hz or so.
Need to filter like crazy to get rid of it.
It drowns out just about everything i try to do.
I suspect but can not prove that my geophone
is out of spec and not working at the 1Hz
it is supposed to.
The only adjustment I can make is to center the
mass in the housing. If there is a way to lower the
resonant freq I would sure like to know.
I am thinking of finding some weights to put on
the adjustment screw as much weight as can still be
centered.
But it is my experience that for a spring to
lower its frequency it needs to extend or compress
farther in distance. similar to a pendulums length.
T=2PiSqRt(length/gravity)
|
|
|
| Back to top |
|
|
gmv
Guest
|
| Posted:
Sat Feb 05, 2005 4:41 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
"Watson A.Name - "Watt Sun, the Dark Remover"" <NOSPAM@dslextreme.com> wrote in message news:11094u4911pj497@corp.supernews.com...
| Quote: |
"gmv" <noemail@please.123> wrote in message
news:CAQMd.48480$EG1.21315@attbi_s53...
I have a fear in dealing with oscillators
I have no ham license and do not want to
generate more RF energy then currently exists.
I choose to go passive and look only at
baseband signals.
If I wanted to do things right i would
use a capacitive sensor in a kind of
bridge configuration and the capacitor would
FM or AM a carrier frequency
to be later demodulated and amplified.
But like I said there already is too much
RFI and I just do not want to deal with the FCC.
This country might use such an excuse to
throw me back in the nuthouse for the fifth time.
Last summer I built a proximity detector that ran at about 450 to 500
kHz. I was disappointed at its sensitivity, only worked to about 3/4"(2
cm). It radiated some RF, but at that freq. the AM radio could only
pick it up at arm's length. This sensor used a capacitive plate to damp
out the oscillator, but the circuit could be changed to vary the freq as
the capacitance changed. In other words, it could have made a good
position sensor.
|
I figure if you want to measur small capacitances you need a
50 MHz oscillator that will run free and dependent upon
the C of the sensor which is probably going to be in the
pico range.
With that you might interfere with communications
if you are not careful shielding things.
| Quote: |
Anyway, what you need to do is use a mechanical filter, instead of
trying to make an electronic filter. One thought comes to mind. Use a
speaker and microphone coupled by a mechanical link, and add weight to
that link so that it would only respond to freqs below some low value,
such as 1 Hz. This also makes it a bandpass filter.
|
That sounds quite a bit more complex to me
then what I am already doing quite successfully.
What i really need to know is whether or not
to trust a capacitance meter that tells me
different readings for the very same cap
when I change frequencies.
When I change from 80 Hz to 8Hz and
test I get a higher capacitance at 8Hz then
I got at 80 Hz but this change does not
exist when I test non-electrolytic capacitors.
I just built a circuit where the C and R was
tested with my meter and the results came out
nearly exactly as designed so i suspect the
readings at the lower frequencies are more correct
for the readings you need to read an Electrolytic
capacitance near the same frequency you intend to use
it at.
What i really need is to buy some non-electrolytic caps
that I can use in my breadboard but this country does
not seem to sell such a thing at least not to the
general public.
I found the caps I need from China I think
but they only manufacture things in bulk
for industry. I think they will not deal
with the little guy like myself.
I really really hate the fact we have sent a lot of our
industry overseas.
| Quote: |
"Watson A.Name - "Watt Sun, the Dark Remover"" <NOSPAM@dslextreme.com
wrote in message news:1106mgko9mlpiac@corp.supernews.com...
"gmv" <noemail@please.123> wrote in message
news:%nIMd.48003$EG1.44889@attbi_s53...
They dont work at freqs below 10Hz
unless they are very large.
I have never seen a 1/20Hz crystal before.
You're right, a xtal would have to be huge to get down to that freq.
However you could multiply the low freq many times and then filter
it.
"loedown" <zero242@internode.on.net> wrote in message
news:42034d8b$1@duster.adelaide.on.net...
If you need really closely matched bandpass filters, have you
considered crystal filters. I know very little about them, but I do
know they have a very sharp cutoff / roll off
Paul
|
|
|
| Back to top |
|
|
gmv
Guest
|
| Posted:
Sat Feb 05, 2005 4:47 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
The important thing for me is that both caps in the circuit ride at the same temp
so they change in a balanced predictable way.
My circuit is kept inside out of the weather at room temp.
Between 75F and 85F.
"Watson A.Name - "Watt Sun, the Dark Remover"" <NOSPAM@dslextreme.com> wrote in message news:11095vroouhm271@corp.supernews.com...
| Quote: |
"Jim Backus" <jhb@nospam.co.uk> wrote in message
news:TpquPuPd0tCd-pn2-SgulLeF9nLcd@localhost...
On Thu, 3 Feb 2005 21:37:45 UTC, "gmv" <noemail@please.123> wrote:
I am trying to get a reading
within 1% of reality.
I'm not sure reality allows measuring electrolytics to 1% :-)
I've never considered electrolytics to be stable components. Their
typical manufacturing tolerance is -20% +50%.
--
Jim Backus OS/2 user since 1994
But those ceramic caps that are .1 uF or more have tempcos that seem to
make them great temp sensors. I put my fingers on a cold one and it
changes by ten percent or more. Makes me think that I could use one in
a chirping cricket circuit and just like a cricket, it will speed up as
it gets warmer.
|
|
|
| Back to top |
|
|
gmv
Guest
|
| Posted:
Sat Feb 05, 2005 5:03 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
"Watson A.Name - "Watt Sun, the Dark Remover"" <NOSPAM@dslextreme.com> wrote in message news:11097hnlc4n3q25@corp.supernews.com...
| Quote: |
"gmv" <noemail@please.123> wrote in message
news:f_QMd.49525$eT5.27123@attbi_s51...
Sorry, Do not know what a DSP/FIR is.
I am using only 8 bit data.
Surprising enough the filters work quite well
between 1Hz and DC just like they do between
1Hz and infinity. The only problem I have is
getting those capacitors in a non-electrolytic form.
I have found I can get pretty good matches in frequency
simply if I can know the true value of the capacitors.
The design matches the theory if only I can
match the capacitors.
I used to laugh at precision and accuracy but
I have found that is the only way to go when
dealing with filter circuits and possibly
oscillator circuits too.
The only way I will play with things like FIR or
DSP is if it will handle data from a non-linearI still think
kind of sensor that tends to overdrive the
preamp at high freqs making it impossible to
look at the low freqs also it got to be cheap.
I need to filter out
the upper freqs right off the bat or it is
impossible to get enough gain to look at the low.
I still think you need to raise the impedance of the circuits
up high enough so that the capacitors are much smaller.
|
That increases the noise.
| Quote: |
Using FET input opamps and megohm resistors, you should be able
to get down to very low freqs with caps of a microfarad.
|
I think I need BJT types if I remember right
they are lower in noise. I have found chopper
amps to be out of the question.
| Quote: |
I offered that Action Electronics web page with the motor capacitors,
but I don't know if they are electrolytic or not. Maybe someone
could help. They also have tantalom caps, which are more stable
than 'lytics. http://www.action-electronics.com/capacitr.htm
|
motor caps are too high power for my needs
I am just using small signal here.
NP caps are what i need.
I think the ceramic ones would be perfect.
I just want ones that will read consistant with my C meter.
Thanks to everyone for your answers
but like always I must seek a solution myself
because it is so rare I really find any answers
on these newsgroups.
It is as if the Engineers and Scientists
avoid this internet stuff like it is the plague.
I just tried building my own cap and a monster
size it was and only 3.3nf...they must be very
hard to make at 10uf or higher and be small.
You need to use the same technology that makes
ICs. Thin layer of metal, thin layer of insulator,
do this repeatedly until you get the right value.
Surely there must be a breakthrough in technology
to make decent high value caps.
Sort of like the movie THIS ISLAND EARTH
I need to find that engineers supplier of
interocitor parts.
| Quote: |
Yes, a log amp is a possibility, I have already
considered that. But the signals I now receive
are good enough for myself and all I wish to do
is to bring closure to my circuit be acquiring what
I consider to be decent capacitors. These electrolytics
are seemingly only for one frequency because
funny things start happening if you read the
capacitance with different frequencies.
Like I get a lower capacitance at a higher frequency.
I have found no such problems with regular
caps that do not use an electrolytic form.
Your solution would be possible if I had a sensor
resonant at 20 seconds but the Geophone I am using
is resonant at something over 1 Hz.
Geospace has just the device I would like to have
they have a .5 Hz Geophone but they want about
$3000 for one and that is way too high for someone
living on Social Security.
"Rob Gaddi" <rgaddi@bcm.YUMMYSPAMtmc.edu> wrote in message
news:cu0elc$t4c@gazette.corp.bcm.tmc.edu...
While I hate to give my usual answer, have you considered just
digitizing the data and running it through a DSP? You're talking
about some massively slow filters to be trying to make with any
reliable tolerance in analog, and with component values that large
it's not going to be cheap. Plus you can use a linear FIR filter
and not have to worry about the phase distortion you were
mentioning.
If your noise is really swamping the signal, and dynamic range is a
concern, at the low speeds you're talking about you could run
the signal through a log amp before you A/D it.
Lucky for me different phases of Earthquakes
seem to prefer different frequencies.
But my picture will never be as good as the
big boys.
Just too darn noisy at this location.
I get a terrible noise at maybe 15 Hz or so.
Need to filter like crazy to get rid of it.
It drowns out just about everything i try to do.
I suspect but can not prove that my geophone
is out of spec and not working at the 1Hz
it is supposed to.
The only adjustment I can make is to center the
mass in the housing. If there is a way to lower the
resonant freq I would sure like to know.
I am thinking of finding some weights to put on
the adjustment screw as much weight as can still be
centered.
But it is my experience that for a spring to
lower its frequency it needs to extend or compress
farther in distance. similar to a pendulums length.
T=2PiSqRt(length/gravity)
|
|
|
| Back to top |
|
|
Boris Mohar
Guest
|
| Posted:
Sat Feb 05, 2005 6:20 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
On Sat, 05 Feb 2005 11:03:00 GMT, "gmv" <noemail@please.123> wrote:
| Quote: | motor caps are too high power for my needs
I am just using small signal here.
NP caps are what i need.
I think the ceramic ones would be perfect.
I just want ones that will read consistant with my C meter.
At these values ceramic caps have very unstable dielectric. Your filter |
will be all over the place with respect to temperature. The voltage
coefficient does not look good either.
Try here:
http://www.electronicstalk.com/news/eas/eas123.html.
Regards,
Boris Mohar
Got Knock? - see:
Viatrack Printed Circuit Designs (among other things) http://www.viatrack.ca |
|
| Back to top |
|
|
Fred Abse
Guest
|
| Posted:
Sat Feb 05, 2005 6:36 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
On Fri, 04 Feb 2005 20:18:02 +0000, Jim Backus wrote:
| Quote: | I'm not sure reality allows measuring electrolytics to 1% :-)
|
Reality allows measuring electrolytics to better accuracy than that.
Unfortunately, however, it doesn't allow *making* them to such tolerances.
Capacitance can also vary with polarizing voltage.
--
Then there's duct tape ...
(Garrison Keillor) |
|
| Back to top |
|
|
gmv
Guest
|
| Posted:
Sat Feb 05, 2005 7:14 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
Yes, that is the kind of thing I am looking for.
In case you are wondering the thing I want to
use them in is here:
As you can tell I am no expert but i can tell you
this will show Earthquakes in a relatively noisy area.
Just got a big one from the Philippines I think.
The big capacitors on the front end are just to
simulate the mass/housing relationship in the
geophone. They are not a reality as far as
capacitors go.
"Boris Mohar" <borism_-void-_@sympatico.ca> wrote in message news:75e9011tvdtt0kmg1202q713e35udgg5rh@4ax.com...
| Quote: | On Sat, 05 Feb 2005 11:03:00 GMT, "gmv" <noemail@please.123> wrote:
motor caps are too high power for my needs
I am just using small signal here.
NP caps are what i need.
I think the ceramic ones would be perfect.
I just want ones that will read consistant with my C meter.
At these values ceramic caps have very unstable dielectric. Your filter
will be all over the place with respect to temperature. The voltage
coefficient does not look good either.
Try here:
http://www.electronicstalk.com/news/eas/eas123.html.
Regards,
Boris Mohar
Got Knock? - see:
Viatrack Printed Circuit Designs (among other things) http://www.viatrack.ca |
|
|
| Back to top |
|
|
Spehro Pefhany
Guest
|
| Posted:
Sat Feb 05, 2005 7:22 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
On Sat, 05 Feb 2005 07:20:18 -0500, the renowned Boris Mohar
<borism_-void-_@sympatico.ca> wrote:
| Quote: | On Sat, 05 Feb 2005 11:03:00 GMT, "gmv" <noemail@please.123> wrote:
motor caps are too high power for my needs
I am just using small signal here.
NP caps are what i need.
I think the ceramic ones would be perfect.
I just want ones that will read consistant with my C meter.
At these values ceramic caps have very unstable dielectric. Your filter
will be all over the place with respect to temperature. The voltage
coefficient does not look good either.
Try here:
http://www.electronicstalk.com/news/eas/eas123.html.
|
Electrolytics are actually pretty stable wrt temperature, time and
applied voltage- unlike those pesky hi-K ceramics. Less microphonics,
too. If your meter doesn't read them consistently, maybe it's time to
invest $50 or $60 in a new meter?
Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com |
|
| Back to top |
|
|
Watson A.Name - \"Watt Su
Guest
|
| Posted:
Sat Feb 05, 2005 9:41 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
"gmv" <noemail@please.123> wrote in message
news:EJ1Nd.51210$eT5.4698@attbi_s51...
| Quote: |
"Watson A.Name - "Watt Sun, the Dark Remover"" <NOSPAM@dslextreme.com
wrote in message news:11097hnlc4n3q25@corp.supernews.com...
"gmv" <noemail@please.123> wrote in message
news:f_QMd.49525$eT5.27123@attbi_s51...
Sorry, Do not know what a DSP/FIR is.
I am using only 8 bit data.
Surprising enough the filters work quite well
between 1Hz and DC just like they do between
1Hz and infinity. The only problem I have is
getting those capacitors in a non-electrolytic form.
I have found I can get pretty good matches in frequency
simply if I can know the true value of the capacitors.
The design matches the theory if only I can
match the capacitors.
I used to laugh at precision and accuracy but
I have found that is the only way to go when
dealing with filter circuits and possibly
oscillator circuits too.
The only way I will play with things like FIR or
DSP is if it will handle data from a non-linearI still think
kind of sensor that tends to overdrive the
preamp at high freqs making it impossible to
look at the low freqs also it got to be cheap.
I need to filter out
the upper freqs right off the bat or it is
impossible to get enough gain to look at the low.
I still think you need to raise the impedance of the circuits
up high enough so that the capacitors are much smaller.
That increases the noise.
Using FET input opamps and megohm resistors, you should be able
to get down to very low freqs with caps of a microfarad.
I think I need BJT types if I remember right
they are lower in noise. I have found chopper
amps to be out of the question.
I offered that Action Electronics web page with the motor
capacitors,
but I don't know if they are electrolytic or not. Maybe someone
could help. They also have tantalom caps, which are more stable
than 'lytics. http://www.action-electronics.com/capacitr.htm
motor caps are too high power for my needs
I am just using small signal here.
NP caps are what i need.
|
Well, the motor caps are non-polarized. That's what you want.
And they are not electrolytic, I believe, but I'm not certain.
And the power doesn't make any diff, but they are large, and
that might be a problem. As I already mentioned, there are
ceiling fan capacitors at the local hardware store that are
about 10 uF, and are not very large. ANd reasonably priced IIRC.
About noise, you decrease noise by reducing the bandwidth. You
reduce the bandwidth by reducing the load on the tuned circuit.
You do this by increasing the impedance of the load on the tuned
circuit, IOW use a higher resistance, and that means FETs or
other very high impedance amps.
I think you are getting more noise because you are decreasing
the bandwidth by keeping the load the same impedance, but
increasing the tuned circuit impedance. That's backwards
from what you should be doing.
| Quote: | I think the ceramic ones would be perfect.
I just want ones that will read consistant with my C meter.
Thanks to everyone for your answers
but like always I must seek a solution myself
because it is so rare I really find any answers
on these newsgroups.
It is as if the Engineers and Scientists
avoid this internet stuff like it is the plague.
I just tried building my own cap and a monster
size it was and only 3.3nf...they must be very
hard to make at 10uf or higher and be small.
You need to use the same technology that makes
ICs. Thin layer of metal, thin layer of insulator,
do this repeatedly until you get the right value.
Surely there must be a breakthrough in technology
to make decent high value caps.
Sort of like the movie THIS ISLAND EARTH
I need to find that engineers supplier of
interocitor parts.
Yes, a log amp is a possibility, I have already
considered that. But the signals I now receive
are good enough for myself and all I wish to do
is to bring closure to my circuit be acquiring what
I consider to be decent capacitors. These electrolytics
are seemingly only for one frequency because
funny things start happening if you read the
capacitance with different frequencies.
Like I get a lower capacitance at a higher frequency.
I have found no such problems with regular
caps that do not use an electrolytic form.
Your solution would be possible if I had a sensor
resonant at 20 seconds but the Geophone I am using
is resonant at something over 1 Hz.
Geospace has just the device I would like to have
they have a .5 Hz Geophone but they want about
$3000 for one and that is way too high for someone
living on Social Security.
"Rob Gaddi" <rgaddi@bcm.YUMMYSPAMtmc.edu> wrote in message
news:cu0elc$t4c@gazette.corp.bcm.tmc.edu...
While I hate to give my usual answer, have you considered just
digitizing the data and running it through a DSP? You're talking
about some massively slow filters to be trying to make with any
reliable tolerance in analog, and with component values that large
it's not going to be cheap. Plus you can use a linear FIR filter
and not have to worry about the phase distortion you were
mentioning.
If your noise is really swamping the signal, and dynamic range is
a
concern, at the low speeds you're talking about you could run
the signal through a log amp before you A/D it.
Lucky for me different phases of Earthquakes
seem to prefer different frequencies.
But my picture will never be as good as the
big boys.
Just too darn noisy at this location.
I get a terrible noise at maybe 15 Hz or so.
Need to filter like crazy to get rid of it.
It drowns out just about everything i try to do.
I suspect but can not prove that my geophone
is out of spec and not working at the 1Hz
it is supposed to.
The only adjustment I can make is to center the
mass in the housing. If there is a way to lower the
resonant freq I would sure like to know.
I am thinking of finding some weights to put on
the adjustment screw as much weight as can still be
centered.
But it is my experience that for a spring to
lower its frequency it needs to extend or compress
farther in distance. similar to a pendulums length.
T=2PiSqRt(length/gravity)
|
|
|
| Back to top |
|
|
R.Lewis
Guest
|
| Posted:
Sat Feb 05, 2005 9:41 pm Post subject:
Re: Accurately measure C of electrolytic capacitors ? |
|
|
"gmv" <noemail@please.123> wrote in message
news:Y3RMd.19306$C24.13288@attbi_s52...
| Quote: |
"R.Lewis" <h.lewis@connect-2.co.uk> wrote in message
news:36htfrF5289nrU1@individual.net...
"gmv" <noemail@please.123> wrote in message
news:HgIMd.47990$EG1.27201@attbi_s53...
"Watson A.Name - "Watt Sun, the Dark Remover"" <NOSPAM@dslextreme.com
wrote in message news:1106d0pdpsb0i35@corp.supernews.com...
"gmv" <noemail@please.123> wrote in message
news:PczMd.46765$IV5.32350@attbi_s54...
snip
Run it through an A/D to get it into the digital domain where you can
handle
it sensibly without having to consider what temperature, sonics, etc
will
otherwise do to your filters if you need that sort of accuracy.
I must filter out the high stuff first because my sample rate
is only 65539.3 samples per hour and my dynamic range is only
8 bits. Like I told someone else my sensor has a non-linear
response in the frequency range of interest.
2Hz to 1/20 Hz.
It would be so wonderful to have a dead end at 2HZ kind
of filter then most of my troubles would evaporate.
|
In which case why not, as I said before, run it through an AtoD etc etc. and
do a proper job on it.
Simple enough to tailor exactly to your needs. |
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in