Wire sizes for low-current measurements

D

DG

Guest
I have a long-standing debate with someone about the sizes of wire used
in sensitive, low-current measurements. He says that one MUST use very
small wires to make the contact. He says that using larger wire might
lead to a contact problem, or noise induced by tension in the thicker,
and hence stiffer wire. He may have made an argument about leakage
before, but that is moot because our measurements are guarded.

I think it doesn't matter and that the only reason small wires have
traditionally been used in this situations, are that the contacts are
usually very small (ie. bond pads).

In my situation, I have a 1cmx1cm sample, and I have deposited 1mmx1mm
metal contacts on each corner, for 4-probe measurements. Then I want to
attach wire with silver paste to the 4-probe contacts. To me, it
doesn't matter whether I use 22 gauge wire or 28 gauge wire, but he
insists I use the smallest wire possible.

I think that a larger wire might in fact be better, as it has a larger
contact area and less resistance (and hence noise). Although in our
case, the sample resistance is so high that any other resistances are
really irrelevant.

I need some physics reasoning here please.

Thanks,
David
 
In article <c3q65j$5jr$2@rumours.uwaterloo.ca>,
DG <david.grant-NOSPAM@telus.net> wrote:
[... measuring tiny currents ...]
I think it doesn't matter and that the only reason small wires have
traditionally been used in this situations, are that the contacts are
usually very small (ie. bond pads).
Click to expand...
Is heat flow an issue in the experiment? If so, thin wires may be a good
idea.

Are you making a twisted pair to prevent pickup? Thinner wire makes a
tighter pair.

Will mechanical stress change the measured quantity?


--
--
kensmith@rahul.net forging knowledge
 
Subject: Wire sizes for low-current measurements
From: DG david.grant-NOSPAM@telus.net
Date: 3/23/2004 2:20 PM Central Standard Time
Message-id: <c3q65j$5jr$2@rumours.uwaterloo.ca

I have a long-standing debate with someone about the sizes of wire used
in sensitive, low-current measurements. He says that one MUST use very
small wires to make the contact. He says that using larger wire might
lead to a contact problem, or noise induced by tension in the thicker,
and hence stiffer wire. He may have made an argument about leakage
before, but that is moot because our measurements are guarded.

I think it doesn't matter and that the only reason small wires have
traditionally been used in this situations, are that the contacts are
usually very small (ie. bond pads).

In my situation, I have a 1cmx1cm sample, and I have deposited 1mmx1mm
metal contacts on each corner, for 4-probe measurements. Then I want to
attach wire with silver paste to the 4-probe contacts. To me, it
doesn't matter whether I use 22 gauge wire or 28 gauge wire, but he
insists I use the smallest wire possible.

I think that a larger wire might in fact be better, as it has a larger
contact area and less resistance (and hence noise). Although in our
case, the sample resistance is so high that any other resistances are
really irrelevant.

I need some physics reasoning here please.

Thanks,
David
Click to expand...
Hi, David. I'm inferring from your description that you're measuring
resistance. Some considerations affecting wire gage size include:

* For current shunt measurements, a smaller wire on sense means a smaller and
more precise point of contact. If a difference of .020" in the measurement
point makes a measurement difference, using 28AWG instead of 22AWG might be
important.

* If you're making a measurement where changing the temperature of what you're
measuring may be significant, you should be aware that 28 AWG has 1/4 of the
cross-sectional area as 22 AWG, and as such will act like much less of a heat
sink.

* Using silver paste is very helpful in this respect, but there will still be a
greater amount of leaching when soldering when you apply more heat. It's a
good precaution to apply as little heat to the pad as possible. A 28AWG wire
will require less heat.

* The thicker and stiffer the wire, the more chance of shearing off the pad.
A skinnier wire might be more flexible.

* If you're measuring resistance, you should check your meter's manual to
determine maximum series resistance in leads for Kelvin measurements.
Depending on the meter, this could be a consideration against 28AWG wire for
longer wiring runs. Most meters which have 4-terminal Kelvin measurements are
pretty resistant to noise (unless you've got a meter that has high-speed
measurements like 100/sec). If noise is an issue, just use a slower
measurement cycling speed on the meter if possible.

If you're measuring extremely high resistance values, you might have some
problems with triboelectric effect, which is the generation of charge from the
wire rubbing against the insulator. For most resistance measurements below the
megohm range on most meters, this is not significant. But you haven't given
many specifics about your measurement, so a first guess would be that it isn't
very important. If triboelectric effect is a problem, you can get wire with a
lubricant between the conductor and the insulator, which essentially eliminates
this problem. I've found, though, that just tying down the wires to keep them
from moving during measurement is as effective.

None of these are "killers". A lot of these are good reasons for using 28AWG
wire as a default, but I think your best answer would be dependent on the exact
circumstances of your measurements, the instruments you're using, the accuracy
you require in your measurements, and the tradeoffs you're willing to make for
increased accuracy. Usually, though, the answer that's good enough is good
enough. I've done it both ways, and I usually used 22AWG stranded for
convenience, durability, and improved reliability. I've used 28AWG for small
test points, and where I didn't want the heat sinking of the wire to affect the
measurement.

Good luck
Chris
 
the only reason small wires have traditionally been used
...are that the contacts are usually very small
David Grant
Click to expand...
The only reason to use large wires is durability.
e.g. Automotive wiring harnesses don't go below 18AWG
because the wires would be too fragile.

Sensing circuits should be >1Mohm.
Even if the wires were 30AWG nichrome,
the voltage drop in the wires would be of no consequence.
 
Ken Smith wrote:
In article <c3q65j$5jr$2@rumours.uwaterloo.ca>,
DG <david.grant-NOSPAM@telus.net> wrote:
[... measuring tiny currents ...]

I think it doesn't matter and that the only reason small wires have
traditionally been used in this situations, are that the contacts are
usually very small (ie. bond pads).


Is heat flow an issue in the experiment? If so, thin wires may be a good
idea.
Click to expand...
The equipment we are using is sometimes used to do temperature dependant
conductivity or Hall mobility. We will just be using room temperature
however.

Are you making a twisted pair to prevent pickup? Thinner wire makes a
tighter pair.
Click to expand...
The device (a semiconductor sample with 4 probes on it) will be in a
sealed vacuum chamber so I don't think pickup is an issue.

Will mechanical stress change the measured quantity?
Click to expand...
No.

Do my answers change your answers significantly?

Dave
 
CFoley1064 wrote:
Subject: Wire sizes for low-current measurements
From: DG david.grant-NOSPAM@telus.net
Date: 3/23/2004 2:20 PM Central Standard Time
Message-id: <c3q65j$5jr$2@rumours.uwaterloo.ca

I have a long-standing debate with someone about the sizes of wire used
in sensitive, low-current measurements. He says that one MUST use very
small wires to make the contact. He says that using larger wire might
lead to a contact problem, or noise induced by tension in the thicker,
and hence stiffer wire. He may have made an argument about leakage
before, but that is moot because our measurements are guarded.

I think it doesn't matter and that the only reason small wires have
traditionally been used in this situations, are that the contacts are
usually very small (ie. bond pads).

In my situation, I have a 1cmx1cm sample, and I have deposited 1mmx1mm
metal contacts on each corner, for 4-probe measurements. Then I want to
attach wire with silver paste to the 4-probe contacts. To me, it
doesn't matter whether I use 22 gauge wire or 28 gauge wire, but he
insists I use the smallest wire possible.

I think that a larger wire might in fact be better, as it has a larger
contact area and less resistance (and hence noise). Although in our
case, the sample resistance is so high that any other resistances are
really irrelevant.

I need some physics reasoning here please.

Thanks,
David


Hi, David. I'm inferring from your description that you're measuring
resistance. Some considerations affecting wire gage size include:
Click to expand...
Yeah that's right I'm measuring resistance basically. As well as the
Hall voltage.

* For current shunt measurements, a smaller wire on sense means a smaller and
more precise point of contact. If a difference of .020" in the measurement
point makes a measurement difference, using 28AWG instead of 22AWG might be
important.
Click to expand...
The measurement point isn't an issue since I'm contacting the wire to a
large contact.

* If you're making a measurement where changing the temperature of what you're
measuring may be significant, you should be aware that 28 AWG has 1/4 of the
cross-sectional area as 22 AWG, and as such will act like much less of a heat
sink.
Click to expand...
Interesting

* Using silver paste is very helpful in this respect, but there will still be a
greater amount of leaching when soldering when you apply more heat. It's a
good precaution to apply as little heat to the pad as possible. A 28AWG wire
will require less heat.

* The thicker and stiffer the wire, the more chance of shearing off the pad.
A skinnier wire might be more flexible.
Click to expand...
Good point.

* If you're measuring resistance, you should check your meter's manual to
determine maximum series resistance in leads for Kelvin measurements.
Depending on the meter, this could be a consideration against 28AWG wire for
longer wiring runs. Most meters which have 4-terminal Kelvin measurements are
pretty resistant to noise (unless you've got a meter that has high-speed
measurements like 100/sec). If noise is an issue, just use a slower
measurement cycling speed on the meter if possible.

If you're measuring extremely high resistance values, you might have some
problems with triboelectric effect, which is the generation of charge from the
wire rubbing against the insulator. For most resistance measurements below the
megohm range on most meters, this is not significant. But you haven't given
many specifics about your measurement, so a first guess would be that it isn't
very important. If triboelectric effect is a problem, you can get wire with a
lubricant between the conductor and the insulator, which essentially eliminates
this problem. I've found, though, that just tying down the wires to keep them
from moving during measurement is as effective.

None of these are "killers". A lot of these are good reasons for using 28AWG
wire as a default, but I think your best answer would be dependent on the exact
circumstances of your measurements, the instruments you're using, the accuracy
you require in your measurements, and the tradeoffs you're willing to make for
increased accuracy. Usually, though, the answer that's good enough is good
enough. I've done it both ways, and I usually used 22AWG stranded for
convenience, durability, and improved reliability. I've used 28AWG for small
test points, and where I didn't want the heat sinking of the wire to affect the
measurement.
Click to expand...
Does 28AWG come in stranded? Any difference whether I use stranded or
not? Besides mechanical/physical considerations (ie. stranded being
better for soldering and reliability, durability, like you said).

Thanks,
Dave
 
I read in sci.electronics.design that DG <david.grant-NOSPAM@telus.net>
wrote (in <40620151.2060404@telus.net>) about 'Wire sizes for low-
current measurements', on Wed, 24 Mar 2004:
Does 28AWG come in stranded?
You could make it from four strands of 34 AWG.
Click to expand...
--
Regards, John Woodgate, OOO - Own Opinions Only.
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
 
In article <c3sv8f$fl9$2@rumours.uwaterloo.ca>,
DG <david.grant-NOSPAM@telus.net> wrote:
[.. Go back and read it if you need the context ...]
Do my answers change your answers significantly?
Click to expand...
No, I think you can use any sized wire you find handy with no problem.
Since the contact point is about 1mm in size, you shouldn't go above that
because the wire will be hard to attach.

--
--
kensmith@rahul.net forging knowledge
 
In article <40620151.2060404@telus.net>,
DG <david.grant-NOSPAM@telus.net> wrote:
[....]
Does 28AWG come in stranded?
Click to expand...
I've seen #26 stranded so I assume #28 comes stranded.


--
--
kensmith@rahul.net forging knowledge
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top