Searching \ for '[EE:] Sensing very small current changes' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/power.htm?key=current
Search entire site for: 'Sensing very small current changes'.

Exact match. Not showing close matches.
PICList Thread
'[EE:] Sensing very small current changes'
2004\10\25@155452 by Denny Esterline

picon face
I'm looking for recomendations on how to measure a current with
extremely fine precision - nanoamps. The signal I'm interested in will
be at about 50 ma all the time, but I'm interested in sub microamp
drift over a period of several hours. Expect dynamic range to be less
than 50ua.

Sugestions ? - search keywords?

Thanks
-Denny
____________________________________________

2004\10\25@164024 by Spehro Pefhany

picon face
At 03:54 PM 10/25/2004 -0400, you wrote:
>I'm looking for recomendations on how to measure a current with
>extremely fine precision - nanoamps. The signal I'm interested in will
>be at about 50 ma all the time, but I'm interested in sub microamp
>drift over a period of several hours. Expect dynamic range to be less
>than 50ua.
>
>Sugestions ? - search keywords?

You might look at using a 24 bit ADC with some ultra-stable resistors.
If you need to create a virtual ground, you'll need to add some additional
complexity- a precision high-gain op-amp and a buffer, most likely.

1nA represents about 0.02ppm, so any kind of temperature or other drift in
your reference or measurement circuit will quickly swamp that. At 100nA
you have a fighting chance without getting into really exotic stuff, but
you're still essentially talking about building a 6 digit DVM.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
spam_OUTspeffTakeThisOuTspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com




____________________________________________

2004\10\25@172755 by olin_piclist

face picon face
Denny Esterline wrote:
> I'm looking for recomendations on how to measure a current with
> extremely fine precision - nanoamps. The signal I'm interested in will
> be at about 50 ma all the time, but I'm interested in sub microamp
> drift over a period of several hours. Expect dynamic range to be less
> than 50ua.

This is going to be VERY difficult.  50nA resolution out of 50mA is one part
in a million, or 1 ppm, or about one count in 20 bits.  If you have to ask,
then this is way over your head.  Forget about it.

Think about this.  Even if you could design a circuit with a temperature
coefficient of 10ppm/degC, your desired output resolution represents only
1/10 deg C temperature change in the circuit.

This sounds to me like the wrong question was asked to begin with.  What is
the overall purpose?  What will be producing current with so little drift?
What are you ultimately trying to measure?


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com
____________________________________________

2004\10\25@173129 by Larry Bradley

flavicon
face
At 04:54 PM 10/25/2004 -0400, you wrote:
 ...
>You might look at using a 24 bit ADC with some ultra-stable resistors.
>If you need to create a virtual ground, you'll need to add some additional
>complexity- a precision high-gain op-amp and a buffer, most likely.

Spehro, where can a hobbyist get high stability high precision (e.g. 0.1%
or better) resistors? I've had a couple of projects in which I needed 0.1%,
but haven't been able to find a source. One of the surplus joints
advertised 0.1% stuff, but when it arrived, it was just 1%. I complained,
they cancelled the bill, but they are still advertising the same stuff
erroneously.

>Best regards,
>
>Spehro Pefhany --"it's the network..."            "The Journey is the reward"
>.....speffKILLspamspam@spam@interlog.com             Info for manufacturers: http://www.trexon.com
>Embedded software/hardware/analog  Info for designers:  http://www.speff.com
>
>
>
>
>_____________________________________________

2004\10\25@180243 by Support - KF4HAZ

flavicon
face
Ever heard of resistor matching?
In a divider (x/2) it is more important that the resistors are equal value than exact value.
A good ohm meter or a bridge circuit are 2 methods,
another is to place them in a divider circuit and read voltages across them.
When all else fails, I use 15 turn trim pots.
It all depends on the application, I have seen precision resistors being used with op-amps that had larger errors than the resistors, 1% or 2% or even 5% parts will do a better job if the offsets are properly nulled and results scaled to precision.

$.02 - KF4HAZ - Lonnie

----- From: "Larry Bradley" <lhbradley@

{Quote hidden}

> >_______________________________________________

2004\10\25@183806 by Jose Da Silva

flavicon
face
On Monday 25 October 2004 02:31 pm, Larry Bradley wrote:
> At 04:54 PM 10/25/2004 -0400, you wrote:
>   ...
>
> >You might look at using a 24 bit ADC with some ultra-stable resistors.
> >If you need to create a virtual ground, you'll need to add some
> > additional complexity- a precision high-gain op-amp and a buffer, most
> > likely.
>
> Spehro, where can a hobbyist get high stability high precision (e.g. 0.1%
> or better) resistors? I've had a couple of projects in which I needed
> 0.1%, but haven't been able to find a source. One of the surplus joints
> advertised 0.1% stuff, but when it arrived, it was just 1%. I complained,
> they cancelled the bill, but they are still advertising the same stuff
> erroneously.

Sounds like a 1-of-a-kind type of project.  :-)
If you are using surface mount resistors, you may want to place several pads
in parallel, then "build" your accurate resistor that way.

Suppose you want a 1000 ohm 0.1% resistor.
Place 11 pads in parallel, then solder 9 10,000 ohm 1% resistors on 9 of the
pads. If you need 9990 ohms to make it equal 1000 ohms 0.1%, then you can
either find yourself a 9990 ohm 1% resistor, or make up the 9990 ohms by
putting 2 resistors in parallel to equal 9990 ohms. Likewise, if you need
10100 to make the parallel resistance equal 1000, then either get a 10100 1%
resistor, or create one by selecting the appropriate parallel resistances to
make 10100 ohms.
Looking at it this way, you're also spreading the generated heat over 10
parts rather than have it go all through 1 part.
____________________________________________

2004\10\25@185938 by Spehro Pefhany

picon face
At 05:31 PM 10/25/2004 -0400, you wrote:
>At 04:54 PM 10/25/2004 -0400, you wrote:
>  ...
>>You might look at using a 24 bit ADC with some ultra-stable resistors.
>>If you need to create a virtual ground, you'll need to add some additional
>>complexity- a precision high-gain op-amp and a buffer, most likely.
>
>Spehro, where can a hobbyist get high stability high precision (e.g. 0.1%
>or better) resistors? I've had a couple of projects in which I needed
>0.1%, but haven't been able to find a source. One of the surplus joints
>advertised 0.1% stuff, but when it arrived, it was just 1%. I complained,
>they cancelled the bill, but they are still advertising the same stuff
>erroneously.

Digikey carries 0.1% 25ppm/K thin film chip resistors at reasonable prices
(like ~0.60 each in 10's, and 13cents each in 1K).

If you want higher stability parts than that- such as wirewound parts- there
are apparently a couple of places in the SFO (?) area that make them up for a
few dollars each (to order). Pease has mentioned them.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
.....speffKILLspamspam.....interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com




____________________________________________

2004\10\25@190023 by Larry Bradley

flavicon
face
Problem with this solution is that you need a way to measure the resulting
resistance to 0.1%

At 03:41 PM 10/25/2004 -0700, you wrote:
{Quote hidden}

Larry Bradley
Orleans (Ottawa), Ontario, CANADA
____________________________________________

2004\10\25@190450 by David Minkler

flavicon
face
In the context of the original post, the issue is much larger than this.

As Olin has correctly pointed out, 50nA in 50mA is 1ppm.  Caddock (one
of the better sources for low TC resistors and networks) produce
networks with TCs as low as 2ppm.  They use the same glop (or film) on
the same substrate and they can't routinely do substantially better than
this.  Add to this the fact that thermoelectric effects are likely to
swamp out your best efforts unless you are very careful.

Now, on the slightly brighter side, perhaps you can rent an instrument
capable of the kind of precision you need.  If this is a one time
measurement, Telogy  < http://www.telogyinc.com > rents the Agilent
3458A 8 1/2 digit DMM.  Yes, it's going to be expensive to rent, but
beats the socks off of anything you are going to be able to build for
less than the full price of one of these puppies.  As to whether or not
the DC current measurement specs meet your needs, you'll have to decide
(it's not 8 1/2 digits worth).

Olin may be right in asking what the real purpose of the measurement is.

Dave

Falcon Wireless Tech Support - KF4HAZ wrote:

{Quote hidden}

>>>_______________________________________________

2004\10\25@190608 by Larry Bradley

flavicon
face
I've done that - gone through piles of 1K 1% resistors and matched them up.
It works when you want matched pairs. It's harder if you need other
ratios.  If I had some accurate resistors, I could use them in a bridge to
match other sets.

You are right - it depends on the application. If you have some method of
calibrating your system, then trim pots (or adjusting the calibration in
the PIC program) works well.

One of the projects I wanted precision resistors for was to build a
precision voltage reference with various outputs - that would solve most of
the problems.

At 05:03 PM 10/25/2004 -0500, you wrote:
{Quote hidden}

2004\10\25@200001 by p.cousens

flavicon
I think olin said it all.
But on the subject of resistors there are many ways to skin the cat.
Say you wanted a 2k1 resistor
2k2,470k and 2m7 in parallel would do it

its easier to adjust to exact values IMHO using parallel

PC

{Original Message removed}

2004\10\25@205153 by Jose Da Silva

flavicon
face
On Monday 25 October 2004 04:00 pm, Larry Bradley wrote:
> Problem with this solution is that you need a way to measure the resulting
> resistance to 0.1%

In that case, don't solder the resistors yet.

let's suppose you bought 1% parts.
Let's suppose that you have a 6 digit ohmmeter, but you don't trust digit
number 6 anyways, so we call it a 5 digit ohmmeter with a nice 6th digit.

Select 10 resistors, measure separately.
10,005.6
10,003.3
10,099.4
09,993.7
10,024.1
09.500,4 <-throw-out, we can't use it
10,042.1
09,929.1
10,110.1 <-throw out, we can't use it
09,971.1
09,991.5
09,940.3
(here you are accurate to 1 ohm assuming you trust your 6digit ohmmeter to 5
digits. or 10,001 to 9999 accuracy)

If all your resistors measure slightly too high, you may want to add an 11th
resistor to bring the value down (let's say resistor #11 = 500k resistor)
If your resistors all measure a little low, you may want to replace the 10th
resistor with a slightly higher value.
Simply put, mix-n-match until the value is as close as you can get.


now for the math:

1/[ (1/10,005) + (1/10,003) + (1/10,099) + (1/9,993) + (1/10,024) +
(1/10,042) + (1/9,929) + (1/9,971) + (1/9,991) + (1/9,940) ]

= 999.948 (expected value)

Now to see how "bad" could it be... supposing we replace digit #5:

worst case (add 1 ohm)
1/[ (1/10,006) + (1/10,004) + (1/10,100) + (1/9,994) + (1/10,025) +
(1/10,043) + (1/9,930) + (1/9,972) + (1/9,992) + (1/9,941) ]
= 1000.048

worst case (subtract 1 ohm)
1/[ (1/10,004) + (1/10,002) + (1/10,098) + (1/9,992) + (1/10,023) +
(1/10,041) + (1/9,928) + (1/9,970) + (1/9,990) + (1/9,939) ]
= 999.848

In summary,
Here you can trust the value to be:
999.948 (-error of 0.1 ohms) or (+error of 0.1 ohms)

According to your ohmmeter, you started with 1 ohm accuracy, but now you know
you have 0.1 ohms accuracy.
You may want to throw out one of the lower resistances such as 9929 and
replace it with the 10110 to bring your 999.948 even closer to 1000.

With the 1% resistors (10100...9900) you bought, and using a 6 digit ohmmeter
which you might trust to let's say 5 digits, you have effectively created a
1000 resistor that is accurate to 0.1 ohms or 0.01%

...plus you've spread heat disipation across 10 parts instead of carrying it
across 1.


{Quote hidden}

____________________________________________

2004\10\25@213039 by Richard.Prosser

flavicon
face

The other thing to watch is temperature stability.

5% resistors  may be about 100 ppm per C, 1%may be 50ppm   &  0.1% may be
30ppm. You will need to make sure that any combination of resistors is
stable "enough" over the operating range required.

RP



On Monday 25 October 2004 04:00 pm, Larry Bradley wrote:
> Problem with this solution is that you need a way to measure the
resulting
> resistance to 0.1%

In that case, don't solder the resistors yet.

let's suppose you bought 1% parts.
Let's suppose that you have a 6 digit ohmmeter, but you don't trust digit
number 6 anyways, so we call it a 5 digit ohmmeter with a nice 6th digit.

Select 10 resistors, measure separately.
10,005.6
10,003.3
10,099.4
09,993.7
10,024.1
09.500,4 <-throw-out, we can't use it
10,042.1
09,929.1
10,110.1 <-throw out, we can't use it
09,971.1
09,991.5
09,940.3
(here you are accurate to 1 ohm assuming you trust your 6digit ohmmeter to
5
digits. or 10,001 to 9999 accuracy)

snip...........





____________________________________________

2004\10\25@221218 by Denny Esterline

picon face
> > I'm looking for recomendations on how to measure a current with
> > extremely fine precision - nanoamps. The signal I'm interested in will
> > be at about 50 ma all the time, but I'm interested in sub microamp
> > drift over a period of several hours. Expect dynamic range to be less
> > than 50ua.
>
> This is going to be VERY difficult.  50nA resolution out of 50mA is one
part
> in a million, or 1 ppm, or about one count in 20 bits.  If you have to
ask,
> then this is way over your head.  Forget about it.

Ahh, but you misunderstood the question (or maybe I phrased it poorly :),
I'm looking for 10nA resolution with a 50uA dynamic range riding on a 50mA
common mode signal. So that's only 1 part in five thousand, about 12 bits -
assuming I can find a way around the common mode current.

> Think about this.  Even if you could design a circuit with a temperature
> coefficient of 10ppm/degC, your desired output resolution represents only
> 1/10 deg C temperature change in the circuit.

That I had realized, and I fully expected to have to put the entire
aparatus into a temperature - presure - humidity stabilized, magneticly
shielded test chamber.

> This sounds to me like the wrong question was asked to begin with.  What
is
> the overall purpose?  What will be producing current with so little
drift?
> What are you ultimately trying to measure?

That's a fair observation and something I thought about after posting. This
list definitely works best when looking at the whole problem instead of one
little detail, so let's start this again.

This is a one off hobby science project - it has (almost) no budget -
consequences of failure is nothing. I want to measure the force of gravity
with enough precision to observe the effects of the moon and sun.

Research has lead me to understand the effects of the moon and sun are
about 200 micro GALs (GAL - named after Galileo, equals 1cm per second
squared, normal gravity about 980 GALs) That would suggest that I need an
instrument that can resolve ~10 uGAL or better.

I've found four types of apparatus so far:
1. Falling object - measure velocity
2. Swinging pendulum - measure time
3. Sensitive spring balance - measure displacement
4. Electro-magnetic balance - measure current

The challenges of each are:
1. As I understand it, these are the state of the art in gravity
measurement, with professional instruments having resolution below 1 uGAL.
The seem to be using laser inferometry to measure the falling object, this
is beyond my budget. However, my math has indicated that I should be able
to achieve 10 uGAL resolution with a 1 GHz frequency source and a couple
inches between optical sensors. 1 GHz counters are a challenge, but doable.
But to produce meaningful results the frequency source would have to have
long and short term stability better than 20PPB (yeah, that's billion,
10^9) That kind of frequency / stability requirement puts into the realm of
atomic standards - and out of my budget.

2. There's a lot of interesting physics to a pendulum, but -basically- in
this context it's a continuous falling object. It's simpler than #1 because
you can count 10s or 100s of cycles. Meaningful results could be attained
with a much slower timebase, perhaps 10 - 100 MHz. But the long term drift
problems remain. I believe this is attainable with a GPS disciplined Oven
Controlled Crystal (OCXO). New this is still way out of the budget, but it
looks like one could be had on e-bay for ~$100 (more than I wanted to
spend, but maybe...), but would I trust it without calibration...

3. Mechanical springs seem like they would be very challenging with lot's
of unquantifiable potential error sources (drift over time / tempco / etc)
but a permanent magnet spring could be done. I found one such project on
the web. The author had used a handmade opto pair to measure the
displacement. He cited the drawbacks as being *very* fiddley to adjust and
very sensitive to variations in the ambient magnetic field (he could 'see'
a small ceramic magnet when it was flipped over ~15 feet away)

4. Electromagnetic balances use an electromagnet to suspend a fixed mass at
a carefully controlled position and then measure the current through the
electromagnet. As I understand it these are normally run under cryogenic
conditions and can resolve 0.01 uGAL. Since I'm only looking for ~10uGAL I
think this is doable.

As a simple test, I hacked up a small speaker. I cut away most of the paper
and support baffle, placed a mass (about 12g) on it and setup a slot pair
and feedback amplifier to control the position. To test it's resolution I
added a small mass (about 80 mg) and I could see a ~1.5 ma change in
current. That works out to about 1nA per uGAL - thus I was looking for a
way to measure current.

So, there's the whole story - any better suggestions?

Thanks
-Denny


____________________________________________

2004\10\26@000117 by Robert Rolf

picon face
Denny Esterline wrote:
> 4. Electromagnetic balances use an electromagnet to suspend a fixed mass at
> a carefully controlled position and then measure the current through the
> electromagnet. As I understand it these are normally run under cryogenic
> conditions and can resolve 0.01 uGAL. Since I'm only looking for ~10uGAL I
> think this is doable.

So you've considered using a sensitive meter movement for your balance?
Instead of cryogenic conditions, why not vacuum since there
will be no convective or radiation pressure effects (in the dark)?

> As a simple test, I hacked up a small speaker. I cut away most of the paper
> and support baffle, placed a mass (about 12g) on it and setup a slot pair
> and feedback amplifier to control the position. To test it's resolution I
> added a small mass (about 80 mg) and I could see a ~1.5 ma change in
> current. That works out to about 1nA per uGAL - thus I was looking for a
> way to measure current.
>
> So, there's the whole story - any better suggestions?

You might want to consider using a laser pen as a distance
measuring interferometer (mirror on balance arm and beam splitter).
It would get you a much more stable control position
since minute displacements would get you walking fringes
that you pick up with a pair of phototransistors.
This would eliminate any DC drift from your
slot pair detector. As a minimum I would suggest using a
differential slot pair detector to reduce temperature effects.
(One diffuse emitter with a detector on either side of your vane).

There are web pages making your own microgram balances using
meter movements, and using a laser pointer pen as an interferometer
over short distances (since it's coherence length is quite small).

Robert


____________________________________________

2004\10\26@024831 by Jake Anderson

flavicon
face
possibly easier than an interferomiter is to use the gross position of the
laser spot on a distant wall. use a CCD camera to look at it.
nowhere near as cool of course ;->

> {Original Message removed}

2004\10\26@030051 by Jose Da Silva

flavicon
face
On Monday 25 October 2004 06:31 pm, @spam@Richard.ProsserKILLspamspampowerware.com wrote:
> The other thing to watch is temperature stability.
>
> 5% resistors  may be about 100 ppm per C, 1%may be 50ppm   &  0.1% may be
> 30ppm. You will need to make sure that any combination of resistors is
> stable "enough" over the operating range required.

So...
select between both carbon based and metalic film resistors.... therefore you
are now probably limited to through-hole parts instead of surface mount
parts since carbon may probably only be available in through hole....

...and
...perhaps throw the whole assembly into a temperature controlled box running
at 30degrees C or some temp a few degrees higher than room temperature.  ;-)

Room temp cools it down, heater warms it up.... could probably hold it within
1degree..... similar idea to temp-controlled crystal ovens.  ;-)


It appears another thread shows a much better description of "the project"
____________________________________________

2004\10\26@055950 by Alan B. Pearce

face picon face
>But to produce meaningful results the frequency source
>would have to have long and short term stability better
>than 20PPB (yeah, that's billion, 10^9) That kind of
>frequency / stability requirement puts into the realm
>of atomic standards - and out of my budget.

Not if you get someone else to pay for the atomic standard - see below.

>Meaningful results could be attained with a much slower
>timebase, perhaps 10 - 100 MHz. But the long term drift
>problems remain. I believe this is attainable with a GPS
>disciplined Oven Controlled Crystal (OCXO). New this is
>still way out of the budget, but it looks like one could
>be had on e-bay for ~$100 (more than I wanted to spend,
>but maybe...), but would I trust it without calibration...

Check out the GPS disciplined oscillator at http://www.rt66.com/~shera/ He
claims to get to a point where the aging rate was 6 parts in 10^12 after 6
months, so reaching the accuracy requirement you listed above should be
readily doable. Now he uses an HP ovened oscillator, but I have intentions
on building one using a cheap AT cut 10MHz crystal in an oven built into a
wide mouth thermos flask I bought for the purpose. The crystal has a
turnover point at 40C, so I reckon I should be able to keep the whole lot
pretty stable by ovening it to 40, and GPS discipline on the frequency.

Interestingly I see there is a scheme to throw out using caesium standards
for scheme that is claimed to be about 3 orders of magnitude better. Won't
happen any time real soon, as they are still developing the concept, but
apparently they reckon it is viable.

____________________________________________

2004\10\26@080614 by olin_piclist

face picon face
Jose Da Silva wrote:
> Place 11 pads in parallel, then solder 9 10,000 ohm 1% resistors on 9
> of the pads.

That gets you a 1.11Kohm 1% resistance, which you could have gotten simply
by starting with a 1.11Kohm 1% resistor.  In this case, the closest standard
value is 1.10Kohms.  The rest of your method works pretty much the same way
from there.

What you are apparently trying to point out is that you can finely adjust a
resistance value by paralleling a predetermined number of fixed resistors of
known tolerance.  This is true, but you are making it way to complicated.
Usually you only need two resistors.

The general method is to pick the first resistor to be the smallest standard
value guaranteed to not be smaller than the desired value.  For each
subsequent resistor but the last one, you measure the resistance so far,
then add the minimum parallel value to guarantee the result will not be less
than the desired value.  For the last resistor, you pick the value that
results in the least overall error.  Only two 1% resistors are needed to
make a .1% resistor within the standard value range.

Let's use your example of making a 1000 ohm .1% resistor from standard 1%
values.  The smallest standard 1% value guaranteed to be >= 1000 ohms is
1020 ohms.  This will be in the range of 1010 to 1030 ohms.  That will
require a parallel resistance of 101K to 34.3K to bring down to 1.000K ohms.
Even in the worst case of the first resistor being 1030 ohms, the closest
standard value for the second is 34.0Kohms, which could be as low as 33,660
ohms.  That in parallel with 1030 ohms yields 999.4 ohms, which is within
the desired 999 to 1001 ohm range.

Note however that this assumes you have the ability to measure a resistance
somewhat better than your desired end tolerance.

Also note that this discussion is not relevant to the original post.  The OP
wanted to measure very small drift in a current value over several hours.
He was interested in precision, not accuracy.  The problem for the OP is
dealing with various causes of drift, not so much initial accuracy.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com
____________________________________________

2004\10\26@085728 by olin_piclist

face picon face
Denny Esterline wrote:
> Ahh, but you misunderstood the question (or maybe I phrased it poorly
> :), I'm looking for 10nA resolution with a 50uA dynamic range riding on
> a 50mA common mode signal. So that's only 1 part in five thousand,
> about 12 bits - assuming I can find a way around the common mode
> current.

But again, initial accuracy isn't the problem since you're willing to
subtract off the initial current.  The problem is measurement drift, which
is relative to the whole current.  You are asking for drift less than .2 ppm
over several hours.

> That I had realized, and I fully expected to have to put the entire
> aparatus into a temperature - presure - humidity stabilized, magneticly
> shielded test chamber.

That only accounts for some sources of drift.  But even looking at
temperature drift alone, let's say you get your *whole circuit* to have only
10ppm/degC drift.  That in itself is a significant feat, and won't come
cheap.  You have to keep that circuit within 1/50 degC to not exceed .2 ppm
error.  This is a very significant challange.  It's not too hard to cook up
something that maintains temperature to 1/2 degC, but 1/50 degC is a totally
different issue.

{Quote hidden}

Why not measure the time over a much longer falling distance?  Frequency
(relative time) and distance are probably the two physical quantities that
an amatuer can create or measure to high accuracy.  Your distance doesn't
need to be known that accurately, only well fixed.  The falling time at 1G
is 450mS after 1 meter from dead start, and 1 second for 4.9 meters.  You
probably want to measure shortly after drop, mid drop, and end to cancell
out variations in initial velocity due to the release mechanism.  In any
case, you've got about 1/2 second to work with given reasonable heights.  A
18F PIC with CCP module can capture events down to 100nS, or .2 ppm over 1/2
second.  Temperature controlled crystals at a little exotic, but not out of
reach for a project like this.  Other sources of errors will be harder to
control, like the effects of the residual vacuum, thermal
expansion/contraction of your reference distance, etc.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com
____________________________________________

2004\10\26@152422 by Peter L. Peres

picon face

On Tue, 26 Oct 2004, Jake Anderson wrote:

> possibly easier than an interferomiter is to use the gross position of the
> laser spot on a distant wall. use a CCD camera to look at it.
> nowhere near as cool of course ;->

This is not so far off, you know. The lacoste-romberg gravimeters use a
microscope type of optics to get a mark's position on a scale (thye are
null reading instruments, the microscope reads the null - in a way). Using
a student microscope could take you a fair way in the desired direction,
it being able to magnify 1000x . Then that would be so much less gain to
make up for in the electronics (measure uA instead of nA f.ex.). For the
target I'd suggest a pyrite cristal which has the necessary high contrast
features. You can get one at a gem shop, and chip off a tiny piece to use
as a target.

Peter
____________________________________________

2004\10\26@210611 by Jose Da Silva

flavicon
face
On Tuesday 26 October 2004 05:03 am, Olin Lathrop wrote:
> Jose Da Silva wrote:
--snip--
> The general method is to pick the first resistor to be the smallest
> standard value guaranteed to not be smaller than the desired value.  For
> each subsequent resistor but the last one, you measure the resistance so
> far, then add the minimum parallel value to guarantee the result will not
> be less than the desired value.  For the last resistor, you pick the value
> that results in the least overall error.  Only two 1% resistors are needed
> to make a .1% resistor within the standard value range.

2 resistors will do, but the granularity of your result will be larger plus less likely that you strike identical values.
Suppose you put 2 resistors in parallel.  27ohms + 27ohms (+/-1 ohm)
Your known value is now 13.5 +/- 0.5
Suppose you put 10 resistors in parallel. 135+...+135 ohms (+/-1 ohm)
Your known value is now 13.5 +/- 0.1

I am more likely "able" to reproduce 2 identical resistors of 13.5+/-0.1 using 10 versus using only 2 resistors.

Let's suppose my ohmmeter is accurate to 3 digits....
I don't like dealing with fractions of ohms here, but just for the sake of using the above 13.5 ohm result.
Say we get 27.4 || 27.4 = 13.7 and we are aiming to be at 13.5+/-0.1 so to get within the same window, I need to put another resistor in parallel of 924ohms (ideal) or 1863 just to get within the window.
Now we build a 2nd 13.5 resistor and find, 26.7, so we need to put something at 27.3 to be accurate... let's use 39.1 || 90.5
Can you reasonably say that the two 13.5 resistors are identical once I start pumping power thru them?
Once the temperature starts going up due to the power, I don't think that the 27.4 || 27.4 || 924 has the same resistance as the 27.3 || 39.1 || 90.5 because they are going to have different loads of power, therefore temp difference, therefore resistance differences.

I would think that the "complicated" 10 resistors stands a better chance of being identical if comparing 1 versus ther 2nd, plus the bonus is reduced temperature generated due to reduced power across each resistor....
You are closer to ambient temperature with 10 versus 2.
Likewise, you are closer to ambient temperature with 20 versus 10.

> Note however that this assumes you have the ability to measure a
> resistance somewhat better than your desired end tolerance.

Of course.
Your numbers are only as good as the tools you use to measure them with.
So, if your meter is only good enough to 2 digits, then 2 digits is all you can trust, but again, the precision is more trustworthy. You are closer to the target value if you spread it over 10 versus 2.

> Also note that this discussion is not relevant to the original post.

If you recall (see ###1), someone else asked a question of where to find 0.1% resistors, so if you can't find them, then you can build them.

> The  OP wanted to measure very small drift in a current value over several
> hours. He was interested in precision, not accuracy.

To get back "On Topic"...   ;-)
After reading the solution you suggested in another thread....
I fully agree with you that resistor accuracy is not relevant considering you have your eye set on trying to propose a time-based weight-drop solution.
Let's put the resistor stuff aside.... ;-)

However, a weight-drop apparatus is only able to measure one instance at a time. you got to reset the apparatus, and therefore you have less guarantee that you set it up identical to the 1st condition. You can't measure steadily over hours of time, therefore you may collect some numbers, but those may or may not be enough to correlate/prove/disprove something.
You can't build the apparatus as part of the vacuum chamber... it has to be free-standing within the chamber, due to air crushing of the chamber.
You can't discount for vibrations or random events which can be filtered out because it is a one-shot type measurement.
Finally, you are trading one set of drift problems with another set of drift problems because you are building a 1-of-a-kind apparatus that you can't reference against an identical 1-of-a-kind apparatus. You can't discount voltage, temp, air pressure,or other drifts because you got nothing to knock your drift values against.
If you were able to build a 2nd "reference" to discount drifts, then a weight-drop solution may be the way to go, but how do you propose that?
you can't build a 2nd weight-drop apparatus in the same room as the 1st because they are both affected by gravity.

> The problem for the
> OP is dealing with various causes of drift, not so much initial accuracy.

For myself, I prefer the 1st solution posted (see ###2) and I would consider that a better suggestion since it can be measured over hours, you can filter-out noises like vibrations, plus you can build 2 identical apparatus.
One is your "no load" for reference purposes to discount for various drifts, the second is your "under load" device measured over time.
You build both with common things, such as common power supply input, common A/D, common......etc. so that you can discount drift. ...whatever affects the refence affects the measured. Some things won't be common between both circuits, but you will aim for identical in function.
Taking the example of the 13.5 ohm resistor above....
Is [ 27.4 || 27.4 || 924 ] identical to [ 27.3 || 39.1 || 90.5 ]
...ummm, no, because power is going to go thru the various resistors at different rates, therefore putting them at different temps, therefore giving different resistances.

You'll probably have a better chance of identical with 10 resistors spread across.... even better, perhaps 20 or more in parallel.
.....perhaps the OT resistor stuff may not be so OT after all.

just my 2 cents.

{Quote hidden}

>
{Quote hidden}

___________________________________________

2004\10\26@212150 by Jose Da Silva

flavicon
face
On Tuesday 26 October 2004 11:49 am, Peter L. Peres wrote:
> On Tue, 26 Oct 2004, Jake Anderson wrote:
> > possibly easier than an interferomiter is to use the gross position of
> > the laser spot on a distant wall. use a CCD camera to look at it.
> > nowhere near as cool of course ;->
>
> This is not so far off, you know. The lacoste-romberg gravimeters use a
> microscope type of optics to get a mark's position on a scale (thye are
> null reading instruments, the microscope reads the null - in a way). Using
> a student microscope could take you a fair way in the desired direction,
> it being able to magnify 1000x . Then that would be so much less gain to
> make up for in the electronics (measure uA instead of nA f.ex.). For the
> target I'd suggest a pyrite cristal which has the necessary high contrast
> features. You can get one at a gem shop, and chip off a tiny piece to use
> as a target.

Consider yourself a genius  ;-)

If I built a 10 meter long horizontal arm holding a weight on the end, I
would be looking for small 0.1 mm movements.... no batteries required.  ;-)

____________________________________________

2004\10\27@110642 by Morgan Olsson

flavicon
face
Denny Esterline 04:16 2004-10-26:
>(GAL - named after Galileo, equals 1cm per second squared, normal gravity about 980 GALs)

Doh.  Why complicate the world?  Why not simply use standard SI system unit m/s^2 ?  All thoose ancient and new created units in some countries make reading technical litterature difficult.  And also it is more difficult to calculate with, to see what unit i get when multiplying Gallileo with time...

Example:

Normal gravity about 9,8 m/s^2

Let that work for 2 seconds i get 19,6 m/s speed after 2 s free fall.

/Morgan
--
Morgan Olsson, Kivik, Sweden

____________________________________________

2004\10\27@192127 by Denny Esterline

picon face
> >(GAL - named after Galileo, equals 1cm per second squared, normal gravity about 980 GALs)
>
> Doh.  Why complicate the world?  Why not simply use standard SI system unit m/s^2 ?  All thoose ancient and new created units in some countries make reading technical litterature difficult.  And also it is more difficult to calculate with, to see what unit i get when multiplying Gallileo with time...
>

I have no idea why they chose to use that unit of measurment. I used
it here because all the reference material I've found continues to use
it- don't ask me why. But when in rome....


-Denny
____________________________________________

2004\10\28@181204 by Peter L. Peres

picon face

On Wed, 27 Oct 2004, Denny Esterline wrote:

>>> (GAL - named after Galileo, equals 1cm per second squared, normal
>>> gravity about 980 GALs)
>>
>> Doh.  Why complicate the world?  Why not simply use standard SI system
>> unit m/s^2 ?  All thoose ancient and new created units in some
>> countries make reading technical litterature difficult.  And also it is
>> more difficult to calculate with, to see what unit i get when
>> multiplying Gallileo with time...
>>
>
> I have no idea why they chose to use that unit of measurment. I used
> it here because all the reference material I've found continues to use
> it- don't ask me why. But when in rome....

The gal is the correct unit in the CGS system. The CGS system is deeply
ingrained in certain branches of chemistry and physics for historical
reasons (most relevant books and papers are in CGS units - argh). CGS is
Centimeter, Gram, Second.

Peter
____________________________________________

More... (looser matching)
- Last day of these posts
- In 2004 , 2005 only
- Today
- New search...