Searching \ for '[EE] Capacitors for DC blocking/low frequency high' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: techref.massmind.org/techref/index.htm?key=capacitors+blockinglow
Search entire site for: 'Capacitors for DC blocking/low frequency high'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Capacitors for DC blocking/low frequency high'
2011\03\08@005025 by V G

picon face
The output from my ECG (instrumentation amplifier) has somewhat of a DC bias
to it.

I would like to get rid of that DC bias and apply a low frequency high pass
filter of about 0.3 Hz to it.

However, filters for such extreme values tend to require high ceramic
capacitor values of 1 or 10 uF to avoid impedance issues and I'm SERIOUSLY
having trouble finding these capacitors at a reasonable price anywhere. Or
is there ANOTHER type of capacitor I can use (not ceramic)?

Or is there another way to go about this

2011\03\08@010741 by Bob Blick

face
flavicon
face


On Tue, 08 Mar 2011 00:50 -0500, "V G"  wrote:
> The output from my ECG (instrumentation amplifier) has somewhat of a DC
> bias
> to it.
>
> I would like to get rid of that DC bias and apply a low frequency high
> pass
> filter of about 0.3 Hz to it.
>
> However, filters for such extreme values tend to require high ceramic
> capacitor values of 1 or 10 uF to avoid impedance issues and I'm
> SERIOUSLY
> having trouble finding these capacitors at a reasonable price anywhere.
> Or
> is there ANOTHER type of capacitor I can use (not ceramic)?

You can use electrolytic capacitors. Just make sure you put them in the
right direction or they will conduct.

Cheerful regards,

Bob

-- http://www.fastmail.fm - Does exactly what it says on the tin

2011\03\08@012425 by Jesse Lackey

flavicon
face
....bipolar electrolytics can be used "either way around" and are often used to AC-couple (i.e. block DC) in audio applications.  Nichicon and panasonic make them, and I'm sure many others do too.  Useful to have in the 'junk drawer'.

J


Bob Blick wrote:
{Quote hidden}

>

2011\03\08@013204 by N. T.

picon face
V G wrote:
> The output from my ECG (instrumentation amplifier) has somewhat of a DC bias
> to it.
>
> I would like to get rid of that DC bias and apply a low frequency high pass
> filter of about 0.3 Hz to it.
>

To  filter 0.3 Hz seems not to be good as 0.5 Hz signal is quite realistic

2011\03\08@031425 by V G

picon face
On Tue, Mar 8, 2011 at 1:07 AM, Bob Blick <spam_OUTbobblickTakeThisOuTspamftml.net> wrote:

> You can use electrolytic capacitors. Just make sure you put them in the
> right direction or they will conduct.
>

What do you mean by "right direction"? The signal ranges from -5V to +5V, so
it charges the capacitor the "right" way part of the time, and "wrong" way
part of the time

2011\03\08@082923 by Olin Lathrop

face picon face
V G wrote:
> I would like to get rid of that DC bias and apply a low frequency
> high pass filter of about 0.3 Hz to it.
>
> However, filters for such extreme values tend to require high ceramic
> capacitor values of 1 or 10 uF to avoid impedance issues and I'm
> SERIOUSLY having trouble finding these capacitors at a reasonable
> price anywhere.

These are widely available and cheap.  However, if you expect to walk to
your local electronics store and buy one for 20 cents, you need to adjust
your expectations.  Real electronics stores don't exist anymore.  There
aren't enough electronic hobbyists in one area to make a physical store
worthwhile.  These things therefore only make sense as mail order.

There are a whole bunch cheaply available from Mouser, for example:

<in.mouser.com/Passive-Components/Capacitors/Ceramic-Capacitors/Multi
layer-Ceramic-Capacitors-MLCC-SMD-SMT/_/N-b2cjZscv7?P=1z0wrkrZ1yzt488Z1yzt47
zZ1yzt489Z1yzt483Z1yzt47b&Keyword=capacitor&Ns=Pricing%7c0&FS=True>

The shipping won't be a big deal if you buy a bunch of stuff at once.  You
seem to want to buy one of something for a "reasonable" price.  That's not
how it works.  Learn to batch things up.  Yes, that means planning ahead.
Impulsiveness is expensive.  Either learn to plan or get over it.

Another thing to do is to build up your own stock.  If you're buying a cheap
part (like 0805 resistors or capacitors) don't just buy one.  Buy 10 or 25
or 100 or wherever a good price break is.  If you needed it once, there's a
good chance you'll need it again.  If you're really into electronics, then
it's a good idea to get a assortment of resistors and capacitors.  Get all
the 5% resistors from 1 Ohm to 1 MOhm.  Such assortments are available
pre-packaged.  For some reason, similar capacitor assortments aren't as
common, but of course you can make a assortment yourself anytime.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000

2011\03\08@084740 by V G

picon face
On Tue, Mar 8, 2011 at 8:29 AM, Olin Lathrop <.....olin_piclistKILLspamspam@spam@embedinc.com>wrote:

> These are widely available and cheap.  However, if you expect to walk to
> your local electronics store and buy one for 20 cents, you need to adjust
> your expectations.  Real electronics stores don't exist anymore.  There
> aren't enough electronic hobbyists in one area to make a physical store
> worthwhile.  These things therefore only make sense as mail order.
>

As much as I appreciate the reply, I would like to respectfully point out
the unnecessary rudeness in your response. "Widely available" and "cheap"
are very relative words. What is cheap to some is expensive to others. What
is widely available to some is scarce to others. I *do* expect to walk into
my local electronics store and buy one for not 20 cents, but probably 2 or 3
as I have in the past. I live (for our purposes), in the University in a
modern, densely populated city, with several electronics parts stores within
walking distance. My favourite one sells a wide variety of components for
cheap prices. And you would be incorrect to assume the lack of demand. Every
time I visit the store, there are at least 10 other people there browsing
for components. Real electronics stores still exist. I know of plenty in my
city, but not all of them are in practical distance for a university student
such as myself.

I'm going to visit the store today and see what they sell in terms of
capacitors.



{Quote hidden}

And how does it not work as "reasonable" price? You imply that I don't
already know how to order items in bulk, or to plan ahead. You are also
suggesting that I am impulsive and therefore irresponsible in my shopping
habits. Are these insults really necessary?


{Quote hidden}

Olin, your responses are much appreciated. But I just don't see the
necessity for the rude, insulting, and condescending nature of your replies..
What do you hope to accomplish with this technique? Why do you feel the
necessity to inject subtle insulting remarks in virtually every single one
of your replies? No matter how strongly you believe that the other person
should focus on only the positive aspect, we are all human and at least
slightly hurt and put off by your attitude

2011\03\08@093942 by Olin Lathrop

face picon face
V G wrote:
> I *do*
> expect to walk into my local electronics store and buy one for not 20
> cents, but probably 2 or 3 as I have in the past.

Then go do it.  Why are you asking the list if this is available to you?

> And how does it not work as "reasonable" price? You imply that I don't
> already know how to order items in bulk, or to plan ahead. You are
> also suggesting that I am impulsive and therefore irresponsible in my
> shopping habits. Are these insults really necessary?

If you know these things, then I fail to see why you asked the list.  What
did you expect us to say?  We don't know anything about your local stores.
What you are asking about is cheaply and widely available, and I used Mouser
as a example.

Too often you seem to be trying to use the list to get away with something
or find that mythical free lunch.  This is even more so since you apparently
did know where to find these items, and therefore what they cost and how
available they are.  You checked Mouser, DigiKey, Jameco, Newark, and
others, didn't like the answer, so came here looking for someone to grant
you a personal exemption to the laws of economics?  This is the kind of
attitude I find annoying, and yes you need to be put in your place.

You have gotten better recently, but you still have the habit of asking for
advice, then trying to get away with something else when it's not the answer
you wanted to hear.  For example, you are apparently finally using a
instrumentation amp for the EKG front end, but you kept arguing about
getting away with using a LM324 for way too long despite numerous people
telling you it was a bad idea.  Don't ask for advice if you're not willing
to listen to the answer.  And if you don't, be prepared for people treating
you in a condescending and insulting way.  You've cleaned up your act some,
but you still have some to go.  Respect is earned, and you haven't gotten
there yet, partly due to the large deficit you created when first bursting
onto the scene.  Whining about being mistreated is only making things worse,
not better.

{Quote hidden}

There was nothing insulting there.  I think others would attest to having
your own stock being a good idea.  I have considerable stock myself.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000

2011\03\08@104058 by V G

picon face
On Tue, Mar 8, 2011 at 9:40 AM, Olin Lathrop <olin_piclistspamKILLspamembedinc.com>wrote:

> V G wrote:
> > I *do*
> > expect to walk into my local electronics store and buy one for not 20
> > cents, but probably 2 or 3 as I have in the past.
>
> Then go do it.  Why are you asking the list if this is available to you?
>

If you would re-read the reply, I stated that I would *expect* to find them,
not that I have already found the specific values I'm discussing here.


{Quote hidden}

That's like saying the sun revolves around the Earth because you see it rise
in the East and set in the West.


Too often you seem to be trying to use the list to get away with something
> or find that mythical free lunch.  This is even more so since you
> apparently
> did know where to find these items, and therefore what they cost and how
> available they are.


If you re read the reply as opposed to making an ignorant assumption, I was
suggesting the possibility as a counterargument to your assumption.


> You checked Mouser, DigiKey, Jameco, Newark, and
> others, didn't like the answer, so came here looking for someone to grant
> you a personal exemption to the laws of economics?  This is the kind of
> attitude I find annoying, and yes you need to be put in your place.
>

You completely missed the point yet again due to your primary focus on
being condescending rather than being of assistance.


> You have gotten better recently, but you still have the habit of asking for
> advice, then trying to get away with something else when it's not the
> answer
> you wanted to hear.


If you don't like one answer, go find another. There is no "right answer" to
this question...


{Quote hidden}

If you would again, reread my responses, you would find that my posts are
more "why not?" as opposed to your assumed "I'm right and you're wrong".

There was nothing insulting there.
>

I beg to differ. Check the recent thread made regarding banning you from
posting in newbie threads

2011\03\08@105548 by N. T.

picon face
V G  wrote:
> The output from my ECG (instrumentation amplifier) has somewhat of a DC bias
> to it.
>
> I would like to get rid of that DC bias and apply a low frequency high pass
> filter of about 0.3 Hz to it.
>
> However, filters for such extreme values tend to require high ceramic
> capacitor values of 1 or 10 uF to avoid impedance issues and I'm SERIOUSLY
> having trouble finding these capacitors at a reasonable price anywhere. Or
> is there ANOTHER type of capacitor I can use (not ceramic)?
>
> Or is there another way to go about this?
> --

Yes. You can use LTC2053 instead of LT1168.
LTC2053:
CMRR is typically 116dB.
Average Input Offset Current s typically  1nA.

LT1168
CMRR 90dB Mi

2011\03\08@110117 by Herbert Graf

picon face
On Tue, 2011-03-08 at 08:29 -0500, Olin Lathrop wrote:
> V G wrote:
> > I would like to get rid of that DC bias and apply a low frequency
> > high pass filter of about 0.3 Hz to it.
> >
> > However, filters for such extreme values tend to require high ceramic
> > capacitor values of 1 or 10 uF to avoid impedance issues and I'm
> > SERIOUSLY having trouble finding these capacitors at a reasonable
> > price anywhere.
>
> These are widely available and cheap.  However, if you expect to walk to
> your local electronics store and buy one for 20 cents, you need to adjust
> your expectations.  Real electronics stores don't exist anymore.  There
> aren't enough electronic hobbyists in one area to make a physical store
> worthwhile.  These things therefore only make sense as mail order.

True perhaps in the areas you are thinking of Olin, but very NOT true in
other areas.

VG happens to be in one of the "mechas" for electronics stores: Toronto.

For downtown Toronto try Supremetronic for "regular" electronics stuff
(they are near Spadina and College, there's also a second store that
carries similar stuff on the south side of College just east of
Spadina). You can also hit Active Surplus on Queen, but don't go there
for something specific.

For uptown there are more options on Gordon Baker near Victoria Park and
Steeles. Active Electronics and Sayal are both there.

Sayal is the best without question, they have a few other locations as
well:

http://www.sayal.com/

There are more stores in the west end, but I'm not familiar with that
area so I can't point the OP to any in particular.

Very rarely do I have to place orders at Digikey since my work is very
close to Sayal and Active.

TTYL

2011\03\08@110158 by V G

picon face
On Tue, Mar 8, 2011 at 10:55 AM, N. T. <.....ntypesemiKILLspamspam.....gmail.com> wrote:

> Yes. You can use LTC2053 instead of LT1168.
> LTC2053:
> CMRR is typically 116dB.
> Average Input Offset Current s typically  1nA.
>
> LT1168
> CMRR 90dB Min
>

SOP is going to be a problem, but let me try anyway

2011\03\08@110711 by V G

picon face
n Tue, Mar 8, 2011 at 11:00 AM, Herbert Graf <EraseMEhkgrafspam_OUTspamTakeThisOuTgmail.com> wrote:

> True perhaps in the areas you are thinking of Olin, but very NOT true in
> other areas.
>
> VG happens to be in one of the "mechas" for electronics stores: Toronto.
>
> For downtown Toronto try Supremetronic for "regular" electronics stuff
> (they are near Spadina and College, there's also a second store that
>

I'll check that out today.


> carries similar stuff on the south side of College just east of
> Spadina).


Are you talking about Creatron? That's the one I go to regularly.


> You can also hit Active Surplus on Queen, but don't go there
> for something specific.
>

I'll check that out as well.


> For uptown there are more options on Gordon Baker near Victoria Park and
> Steeles. Active Electronics and Sayal are both there.
>

I've been to the Sayal there but never found the Active Electroincs. Sayal
has a lot of cool stuff but is *VERY* expensive.

Sayal is the best without question, they have a few other locations as
> well:

2011\03\08@111859 by Olin Lathrop

face picon face
N. T. wrote:
> Yes. You can use LTC2053 instead of LT1168.
> LTC2053:
> CMRR is typically 116dB.
> Average Input Offset Current s typically  1nA.
>
> LT1168
> CMRR 90dB Min

How is that supposed to help with a DC bias voltage problem?  What he needs
it high pass filtering, which has nothing to do with the common mode
rejection ratio of the instrumentation amp.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000

2011\03\08@112429 by V G

picon face
On Tue, Mar 8, 2011 at 11:19 AM, Olin Lathrop <olin_piclistspamspam_OUTembedinc.com>wrote:

> N. T. wrote:
> > Yes. You can use LTC2053 instead of LT1168.
> > LTC2053:
> > CMRR is typically 116dB.
> > Average Input Offset Current s typically  1nA.
> >
> > LT1168
> > CMRR 90dB Min
>
> How is that supposed to help with a DC bias voltage problem?  What he needs
> it high pass filtering, which has nothing to do with the common mode
> rejection ratio of the instrumentation amp.


He was most likely addressing the issue of a 60Hz signal going through the
heart signal. I posted about it a little while ago. Perhaps he posted in the
wrong thread

2011\03\08@164902 by IVP

face picon face
>> Olin, your responses are much appreciated. But I just don't see the
>> necessity for the rude, insulting, and condescending nature of your
>> replies.
>
> There was nothing insulting there.  I think others would attest to
> having your own stock being a good idea

I agree. It's commonsense long-term economics

The LCDs I mentioned a couple of days ago ? I need 2 but I'll get 4

Not always to save money on the part itself, but to spread shipping
costs, and the chance of me NOT needing an LCD in the near future
is negligible

Many years ago I bought a box of 100 of every E12 value 1/4W
resistor from 10 to 10M. Still digging in to it, still handy to have. I
did the same last year with 0805 SMT resistors

I also have many boxes of PCBs, recovered from skips and work
shops that have thousands of perfectly serviceable parts (including
high-value SMT ceramics and bipolar electrolytics) waiting to be
scavenged. Our ancient F&P washing machine broke down a few
weeks back. Already I've recovered some LEDs and other (tested)
parts from it to make my solar tracker. The motor will be turned into
a wind generator. I designed a 3-phase PCB for someone a few
years ago to use such motors for that

But I digress


I appreciate for VG there might be space and $ considerations but
it is worthwhile adding to stock whenever possible

Component shopping used to be more fun than inconvenience and
it is very satisfying for me to know now that I've already got 99% of
the bits I need for most projects, either loose or a couple of solder
joints away


Jo

2011\03\09@033236 by N. T.

picon face
Olin Lathrop wrote:
> N. T. wrote:
>> Yes. You can use LTC2053 instead of LT1168.
>> LTC2053:
>> CMRR is typically 116dB.
>> Average Input Offset Current s typically  1nA.
>>
>> LT1168
>> CMRR 90dB Min
>
> How is that supposed to help with a DC bias voltage problem?  What he needs
> it high pass filtering, which has nothing to do with the common mode
> rejection ratio of the instrumentation amp.
>

Yes, I over-truncated the specs, my mistake..
---
The LTC2053 is a high precision instrumentation amplifier. The CMRR is
typically 116dB with a single or dual
5V supply and is independent of gain. The input offset voltage is
guaranteed below 10µV with a temperature
drift of less than 50nV/°C
---
The LT1168 is laser trimmed for very low input offset voltage (40µV
max), drift (0.3µV/°C)
---

VG wrote: "The output from my ECG (instrumentation amplifier) has
somewhat of a DC bias to it." I believe he meant that the mean value
of the output waveform is not close enough to 0. The reason, I think,
is input offset voltage and temperature drift.

LTC2053 input offset voltage is below 10µV with a temperature drift of
less than 50nV/°C
LT1168 input offset voltage is 40µV max, drift is 0.3µV/°C.

That's why I suggested using LTC2053 instead of LT1168.

2011\03\09@072252 by V G

picon face
On Wed, Mar 9, 2011 at 3:32 AM, N. T. <@spam@ntypesemiKILLspamspamgmail.com> wrote:

{Quote hidden}

I mean, for some reason, with the electrodes not connected to anything, I
see a steady DC signal of 2V or so for some reason

2011\03\09@081900 by Olin Lathrop

face picon face
N. T. wrote:
> VG wrote: "The output from my ECG (instrumentation amplifier) has
> somewhat of a DC bias to it." I believe he meant that the mean value
> of the output waveform is not close enough to 0. The reason, I think,
> is input offset voltage and temperature drift.

Probably not, although it would help if the OP would quantify the symptoms
he is seeing.  I thought the leads were picking up a DC offset, probably due
to galvanic action.  That is likely to be many times more than the diff amp
input offset times the gain.

In any case, high pass filtering would take care of the problem.  If he
really wants to do this inside the circuit after the diff amp, then it's
pretty easy since you can chose the impedance.  For example, 100KOhms in
series with the diff amp output followed by 10uF to ground creates a 160mHz
high pass filter.  That should be low enough to not interfere with
legitimate signals, but still low enough impedance to be no trouble inside
the circuit.  This can go directly into the non-inverting input of the next
amp.  Just about anything other than a LM324 will be fine with 100KOhm
impedance input signal.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000

2011\03\09@092538 by N. T.

picon face
V G wrote:
> I mean, for some reason, with the electrodes not connected to anything, I
> see a steady DC signal of 2V or so for some reason.
> --

Are you talking about "Figure 8. Nerve Impulse Amplifier" of
1168fa.pdf (LT1168 datasheet)?
Most probably, as for me, the negative feedback loop is not working

2011\03\09@093111 by N. T.

picon face
Olin Lathrop wrote:
> N. T. wrote:
>> VG wrote: "The output from my ECG (instrumentation amplifier) has
>> somewhat of a DC bias to it." I believe he meant that the mean value
>> of the output waveform is not close enough to 0. The reason, I think,
>> is input offset voltage and temperature drift.
>
> Probably not, although it would help if the OP would quantify the symptoms
> he is seeing.  I thought the leads were picking up a DC offset, probably due
> to galvanic action.  That is likely to be many times more than the diff amp
> input offset times the gain.
>
> In any case, high pass filtering would take care of the problem.

Waving a dead fish under the right phase of the Moon would help better :-)
If "leads were picking up a DC offset, probably due to galvanic
action", the right approach, in my opinion, would be to eliminate
"picking up a DC offset".

2011\03\09@101742 by Olin Lathrop

face picon face
N. T. wrote:
>> In any case, high pass filtering would take care of the problem.
>
> Waving a dead fish under the right phase of the Moon would help
> better :-) If "leads were picking up a DC offset, probably due to
> galvanic
> action", the right approach, in my opinion, would be to eliminate
> "picking up a DC offset".

Good idea, let's see how we'd do that.  We'd need something that blocked DC
but let other stuff thru.  Of course we have to block some things close to
DC, since otherwise we'd have to wait forever for the signal to stabalize.
Ooh, ooh, I know.  We want one of them there low frequency blocking
thingies.  I think the fancy name is "filter".  Let's use that so others
think we're more smarterer.  "Low frequency blocking filter" is kindof
cumbersome to say, so let's flip it around and call it a "high frequency
pass-thru filter".  That's still kindof a mouthful.  How about "high
frequency pass filter", or maybe even just "high pass filter".  Oh right,
that's what I said in the first place.  Gee, who'd have thought!?


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000

2011\03\09@104542 by Mark Rages

face picon face
On Wed, Mar 9, 2011 at 9:18 AM, Olin Lathrop <KILLspamolin_piclistKILLspamspamembedinc.com> wrote:
{Quote hidden}

Clearly, he needs to get four op-amps and build this circuit:

http://electronicdesign.com/article/analog-and-mixed-signal/simple-ac-stop-dc-pass-circuit-uses-four-op-amps61.aspx

(I've posted this before; reposting because I think it's hilarious.)
-- Mark Rages, Engineer
Midwest Telecine LLC
RemoveMEmarkragesTakeThisOuTspammidwesttelecine.com

2011\03\09@104624 by N. T.

picon face
Olin Lathrop wrote:
{Quote hidden}

Not exactly. A a DC offset can be dealt with not only by high pass
filtering. A DC offset, due to the galvanic action, you talked above,
is more or less constant thing, you can compensate it in advance, if
you know the expected value.

A good Precision Instrumentation Amplifier, say LTC2053, typically
have "REF" input. Using a nano-watt microcontroller you can analyze
mean value of the output waveform in time depending on the temperature
and other factors and just preset "REF" input to compensate the DC
offset. I would not call the approach "high pass filtering".

2011\03\09@113115 by Bob Blick

face
flavicon
face
On Wed, 09 Mar 2011 16:31 +0200, "N. T." wrote:

> If "leads were picking up a DC offset, probably due to galvanic
> action", the right approach, in my opinion, would be to eliminate
> "picking up a DC offset".

I agree with that, too. If there is an unwanted signal, DC in this case,
in your circuit, get rid of it as soon as possible. If not before the
first stage, then right after.

Cheerful regards,

Bob

-- http://www.fastmail.fm - Send your email first class

2011\03\09@124939 by V G

picon face
On Wed, Mar 9, 2011 at 11:31 AM, Bob Blick <spamBeGonebobblickspamBeGonespamftml.net> wrote:

> On Wed, 09 Mar 2011 16:31 +0200, "N. T." wrote:
>
> > If "leads were picking up a DC offset, probably due to galvanic
> > action", the right approach, in my opinion, would be to eliminate
> > "picking up a DC offset".
>
> I agree with that, too. If there is an unwanted signal, DC in this case,
> in your circuit, get rid of it as soon as possible. If not before the
> first stage, then right after.
>
>
Yeah, I need to go to Home Depot and get some insulated wire. I'm using
regular wire (twisted together) for (+), (-), and GND.

I'll probably then go to the local hospital and get some "real" electrodes.

A low frequency high pass filter after the first amplifier stage should work
well

2011\03\09@130454 by Olin Lathrop

face picon face
On Wed, Mar 9, 2011 at 11:31 AM, Bob Blick <TakeThisOuTbobblickEraseMEspamspam_OUTftml.net> wrote:
> I agree with that, too. If there is an unwanted signal, DC in this
> case, in your circuit, get rid of it as soon as possible. If not
> before the first stage, then right after.

Generally it's good to get rid of noise as early as possible.  Originally I
thought the OP was planning on put a capacitor in line with the sensor,
which makes some sense.  He apparently wants to put it after the first amp.
That's OK as long as the DC offset is small enough so that it doesn't cause
the signal out of the first amp to clip.  He has since said the DC offset is
2V out of the first amp, but since he refuses to provide a schematic it's
hard to say if that is big or small or if there are other things that should
be considered.

Putting the filter after the first amp has the advantage of allowing you to
pick the impedance and thereby making the filter predictable.  This is a
compelling enough advantage that it might make sense to reduce the gain of
the first amp as necessary to tolerate the DC offset.  Its job is to do the
differential detection, provide some gain, and buffer the signal for easy
processing by downstream circuitry, including the high pass filter.  More
common mode rejection or lower input offset of the first amp is not going to
solve this problem.

2011\03\09@131030 by Olin Lathrop

face picon face
N. T. wrote:
> A DC offset, due to the galvanic action, you talked above,
> is more or less constant thing, you can compensate it in advance, if
> you know the expected value.

If the DC offset is caused by galvanic action (just a guess on my part, but
plausible given the little information we have been given), then it is won't
be constant or predictable.  It will depend on how sweaty the skin is at
each electrode, and can shift easily over tens of seconds as the patient
squirms, etc.

About all you can say is that the galvanic action signal will be lower
frequency than useful heart signals, so that provides a easy way to
discriminate between the two.  No matter how you spin it, that's going to
take a high pass filter.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000

2011\03\09@131032 by V G

picon face
On Wed, Mar 9, 2011 at 1:05 PM, Olin Lathrop <RemoveMEolin_piclistspamTakeThisOuTembedinc.com>wrote:

> Putting the filter after the first amp has the advantage of allowing you to
> pick the impedance and thereby making the filter predictable.  This is a
> compelling enough advantage that it might make sense to reduce the gain of
> the first amp as necessary to tolerate the DC offset.  Its job is to do the
> differential detection, provide some gain, and buffer the signal for easy
> processing by downstream circuitry, including the high pass filter.  More
> common mode rejection or lower input offset of the first amp is not going
> to
> solve this problem.


That's what I was thinking

2011\03\09@131412 by N. T.

picon face
Olin Lathrop wrote:
> He has since said the DC offset is
> 2V out of the first amp, but since he refuses to provide a schematic it's
> hard to say if that is big or small or if there are other things that should
> be considered.
>

I believe the schematic is:
"Figure 8. Nerve Impulse Amplifier" of
http://cds.linear.com/docs/Datasheet/1168fa.pdf

2011\03\09@134106 by N. T.

picon face
Olin Lathrop wrote:
>
> About all you can say is that the galvanic action signal will be lower
> frequency than useful heart signals, so that provides a easy way to
> discriminate between the two.  No matter how you spin it, that's going to
> take a high pass filter.
>

The concept of filtering a system's output is completely different to
the concept of regulating a system's parameter to make the output meet
some criteria.

Some would say these are the same concepts, but I say they are different.

2011\03\09@134134 by V G

picon face
On Wed, Mar 9, 2011 at 1:14 PM, N. T. <ntypesemiEraseMEspam.....gmail.com> wrote:

> I believe the schematic is:
> "Figure 8. Nerve Impulse Amplifier" of
> http://cds.linear.com/docs/Datasheet/1168fa.pdf


It's nothing special. Right now, it's just an LT1168 with a variable
resistor for gain setting hooked up to an oscilloscope

2011\03\09@135324 by V G

picon face
On Wed, Mar 9, 2011 at 1:37 PM, N. T. <EraseMEntypesemispamgmail.com> wrote:

> The concept of filtering a system's output is completely different to
> the concept of regulating a system's parameter to make the output meet
> some criteria.
>
> Some would say these are the same concepts, but I say they are different.


You're right, but in this case, all I want is a clean signal without any DC
offset, so a  high pass filter after the first stage will do nicely.

I was more worried about the capacitor arrangement in all of this. I'll
probably use two tantalums back to bac

2011\03\09@141043 by Sean Breheny

face picon face
It may be possible, though, to find a material which is at least
minimally conductive (enough to pass a signal in to a high impedance
amplifier) but which does not generate much of a"battery" when in
contact with sweat and skin. I believe that this is a key element in
the design of real EKG electrodes. Perhaps VG could get some of these
electrodes from someone at his university or even find a way to buy a
small quantity.

Sean


On Wed, Mar 9, 2011 at 1:10 PM, Olin Lathrop <RemoveMEolin_piclistEraseMEspamEraseMEembedinc.com> wrote:
{Quote hidden}

>

2011\03\10@131631 by N. T.

picon face
V G wrote:
>
> I was more worried about the capacitor arrangement in all of this. I'll
> probably use two tantalums back to back
> --

That's the interesting discussion. I tried but seem to fail again, I
should have been more "diplomatic" in this thread. Perhaps in the
future you may consider the idea of keeping the output of the LTC2053
close to zero by manipulating "REF" input using a nano-watt
microcontroller. The measurements will be the microcontroller's output
to "REF" input

2011\03\10@132705 by Sean Breheny

face picon face
Just a warning that I would not put two Tantalums back to back as they
are much less tolerant of reverse voltage than Aluminum electrolytic
capacitors.

I really think, VG, that you should be able to do what you want very
well with high-ish value ceramic capacitors (say 10uF) and a high
value resistor as a load, making the R*C product several seconds long.

Sean


On Wed, Mar 9, 2011 at 1:53 PM, V G <RemoveMEx.solarwind.xspam_OUTspamKILLspamgmail.com> wrote:
{Quote hidden}

>

2011\03\10@145222 by V G

picon face
On Thu, Mar 10, 2011 at 1:27 PM, Sean Breheny <EraseMEshb7spamspamspamBeGonecornell.edu> wrote:

> Just a warning that I would not put two Tantalums back to back as they
> are much less tolerant of reverse voltage than Aluminum electrolytic
> capacitors.
>
> I really think, VG, that you should be able to do what you want very
> well with high-ish value ceramic capacitors (say 10uF) and a high
> value resistor as a load, making the R*C product several seconds long.


Advice taken. I went out today and bought a few 10uF and 100uF aluminum
electrolytic bipolar caps at the local store. Very convenient. Price is very
reasonable.

The time constant is interesting. Since the time constant has to a certain
value for a certain frequency, is it better to have a high R and low C or
high C and low R

2011\03\10@155247 by IVP

face picon face

> The time constant is interesting. Since the time constant has to a certain
> value for a certain frequency, is it better to have a high R and low C or
> high C and low R?

Large general purpose electrolytic caps tend to have high leakage. They
wouldn't be suitable for long accurate RC times or as filter caps on a nW
battery-powered board for example

Ceramics and other non-polarised - mylar, PE etc - and low-value
electrolytics [including tantalum] would have no or no significant leakage

Joe

2011\03\10@161217 by V G

picon face
On Thu, Mar 10, 2011 at 3:52 PM, IVP <RemoveMEjoecolquittKILLspamspamclear.net.nz> wrote:

> Large general purpose electrolytic caps tend to have high leakage. They
> wouldn't be suitable for long accurate RC times or as filter caps on a nW
> battery-powered board for example
>
> Ceramics and other non-polarised - mylar, PE etc - and low-value
> electrolytics [including tantalum] would have no or no significant leakage
>
>
How would you define large and small? Is 10uF large

2011\03\10@164721 by IVP

face picon face
> How would you define large and small? Is 10uF large?

10uF would be a small electrolytic, a large non-polarised

A leakage figure for a general purpose aluminium electrolytic
< 1000uF might be something like 0.01CV or 3uA, whichever
is greater

Larger values, > 1000uF, might be 3CV, +100uA if > 100V,
which can be several mA

so you work out from that what the leakage will be for the voltage
and uF you use and whether it's significant. For example a high
value resistor, eg several Meg, and a cap that leaks a few uA is
going to be a terrible RC combination, because the small current
available through the resistor is leaking away through the cap, not
charging it

The above applies mostly to aluminium electrolytics. Ceramic,
plastic [and tantalum, bearing in mind uF tolerance] caps are
better suited for timin

2011\03\10@174317 by V G

picon face
On Thu, Mar 10, 2011 at 4:47 PM, IVP <joecolquittSTOPspamspamspam_OUTclear.net.nz> wrote:

{Quote hidden}

Thanks for the info. For my bandpass filter, I used 10uF bipolar aluminum
electrolytic caps and 47K and 120 resistor values. I think those should be
reasonable values

More... (looser matching)
- Last day of these posts
- In 2011 , 2012 only
- Today
- New search...