Wheatstone bridge problem

[Matty F]

I've used a Wheatstone Bridge to measure the resistance of 4 large
solenoids. The readings were: 0.3, 0.28, 0.27, 0.26 ohms.
I measured all of them again and then they were all about 0.20 ohms.
As you can see, the readings have slowly dropped from 0.3 to 0.2 ohms.
Can anyone offer suggestions why the readings keep changing?
I assume that the battery condition is independent of the measurements.

Each time I did a measurement I reconnected the alligator clips to the
terminals and wriggled them around to get a good contact, and checked
the reading three times for each solenoid. Each of those three readings
was the same. i.e. I've done a total of 24 readings.

When I connect the two leads from the Bridge together I get a reading of
0.05 ohms. There appears to be several hundred feet of 14 gauge wire on
each solenoid. Each time I press the check button on the bridge the
meter needle swings slightly to the left of its final position. Each
solenoid weighs about 20kg. There is no way they can be changing
temperature!

 

[Matty F]

Chris wrote:

Hi, Matty. You haven't said specifically what instrument you're using,
so I can't give specific advice. But I can say for sure that
Wheatstone bridges which only have two terminals for Rx aren't made to
measure low resistance. You might be getting self-heating in one of
the bridge's internal resistors, or if there's an active null, battery
voltage may be affecting that. These are the first causes that come to
mind.

I'll have a look at the instrument tomorrow. It's a quality instrument,
and I used an identical one 40 years ago at University. There are only
two terminals. I do suspect self-heating in an internal resistor.
I am being criticized for blaming the meter. Do you have any references
for inaccuracy in measuring low resistance, i.e. around 0.2 ohm?

Why don't you try using a 10 volt or so regulated power supply capable
of cranking an amp or so, a 10 ohm or so resistor, and a voltmeter and
ammeter. Connect the resistor in series with the power supply to give
you an approximately 1 amp source. Put the ammeter in series with the
circuit to measure current. Then put the solenoid in the circuit, and
measure voltage across the solenoid.

The boss has already suggested doing something like that, and is going
to bring his Avometer and other meters tomorrow. We have already put 30
amps through another solenoid, and it seems to work fine. We have a
large battery with a 100 amp meter and can change the current with a
compressible pile of carbon slabs.
All we want to do is to check whether all the solenoids are the same,
and whether they have shorted turns. Measuring the precise resistance is
not important. I was just wondering why it keeps changing.

In use, the solenoids get a variable high current (maybe 30 or more
amps) for a maximum of a minute or so.
According to tables I've just read, 100 feet of 14 AWG is about 0.25
ohms, which is in line with what I get.

 

[Matty F]

Chris wrote:

Hi, Matty. The quality of the instrument doesn't matter if it's being
used to measure something it's not made to do. I'd be fairly confident
that if there are only two terminals, it's probably not made to
accurately measure less than an ohm. So don't worry about blaming the
instrument. From what you've said, that's the first place I'd look,
too.

For the record I have found the following references saying that the
simple Wheatstone Bridge is unsuitable for measuring resistances under 1
ohm:

http://www.coe.uncc.edu/~elsheppa/2155/4-Wheatstone.pdf

Two of the more common types of bridges are the Wheatstone Bridge, which
is used to measure resistances of 1 ohm to 100,000 ohms, and the Kelvin
Double-Bridge, which is used to measure resistances in the range of
..0001 ohms to 1 ohm.

http://www.tpub.com/content/neets/14193/css/14193_94.htm

A Kelvin bridge is recommended for measuring resistances lower than 1 ohm.

http://www.8886.co.uk/forum/viewtopic.php?t=3

There are lots of "Wheatstone Bridge" circuits and the simplest of these
are NOT suitable for measuring low resistances below about 1 Ohm. So for
a typical resistance wire investigation the simple Wheatstone Bridge can
not be used.

http://operations.usace.army.mil/hydro/pdfs/bmp-Transformers.pdf

Test: Winding DC Resistance Measurements
Detects: Broken strands, loose connections, bad tap changer contacts.
Tool: Wheatstone Bridge (1 ohm and greater), Kelvin Bridge
micro-ohmmeter (Less than 1 ohm)

High currents are more difficult to measure accurately, though. A
12VDC 1 amp unregulated wall wart, an LM317 with a heat sink, two
resistors and two 10uF caps will get you the regulated 10 volts. Most
of this stuff might even be in your junkbox or scrounge-able.

I have analogue meters up to 100 amps that seem to work fine, and as
much DC power as anyone could want. There might even be a Kelvin Bridge
around here somewhere. Clearly we need something like that. I am only
trying to compare all the solenoids. I've already measured the pull
force for one solenoid and that's OK.
There's more than a dozen to check.

 

[Matty F]

Chris wrote:

Hi, Matty. A 100 amp analog meter certainly isn't going to be good
enough here. Their accuracy is specified as a percentage of full
scale. Even a very good 100 amp meter might be specified as +/-2%,
which would be +/- 2 amps. You're not going to get there from where
you are. That meter will make you want to push higher current to get a
more accurate reading, which will cause Rx heating and ruin your
measurement.

The same is true for DC current applied. Don't think large -- you'll
ruin your measurement. Think small current, like 1 amp.

I have fed current to the solenoid and measured the current and voltage
with Avometers. The current was 1.45 amps (on the 1.5 amp scale) and the
voltage 0.38 volts, which works out to 0.26 ohms, which is the same as
what the Wheatstone bridge measured initially yesterday, and again
today. I've avoided pressing the test button too much this time.

The Wheatstone brand is "Pontavi" or some name like that. The name has a
strange font so I can't be certain it's right. I cannot see that name on
the Net. I'm told it has a 9 volt battery in it. If the lowest range
used two 5 ohm resistors in the bridge, when measuring a 0.25 ohm
resistance it would be trying to push nearly 2 amps through the bridge
resistors, which could heat them up to have the effect I got before when
the bridge readings kept going down for each measurement.
There is no "sensitivity" switch (which is badly needed).
There is no switch to put power on the resistors before the meter is
used, so the large impedance of the solenoid makes the meter move when
the test button is pressed.

Everybody but me in the workshop says that the Wheatstone bridge is fine
to measure resistances of way under 1 ohm. So clearly you and I and all
the references that I gave before are wrong. All my University work in
physics measurements and statistics was clearly wasted.
Anyway, thanks for your help.

 

[Chris]

Matty F wrote:

I've used a Wheatstone Bridge to measure the resistance of 4 large
solenoids. The readings were: 0.3, 0.28, 0.27, 0.26 ohms.
I measured all of them again and then they were all about 0.20 ohms.
As you can see, the readings have slowly dropped from 0.3 to 0.2 ohms.
Can anyone offer suggestions why the readings keep changing?
I assume that the battery condition is independent of the measurements.

Each time I did a measurement I reconnected the alligator clips to the
terminals and wriggled them around to get a good contact, and checked
the reading three times for each solenoid. Each of those three readings
was the same. i.e. I've done a total of 24 readings.

When I connect the two leads from the Bridge together I get a reading of
0.05 ohms. There appears to be several hundred feet of 14 gauge wire on
each solenoid. Each time I press the check button on the bridge the
meter needle swings slightly to the left of its final position. Each
solenoid weighs about 20kg. There is no way they can be changing
temperature!

Hi, Matty. You haven't said specifically what instrument you're using,
so I can't give specific advice. But I can say for sure that
Wheatstone bridges which only have two terminals for Rx aren't made to
measure low resistance. You might be getting self-heating in one of
the bridge's internal resistors, or if there's an active null, battery
voltage may be affecting that. These are the first causes that come to
mind.

Why don't you try using a 10 volt or so regulated power supply capable
of cranking an amp or so, a 10 ohm or so resistor, and a voltmeter and
ammeter. Connect the resistor in series with the power supply to give
you an approximately 1 amp source. Put the ammeter in series with the
circuit to measure current. Then put the solenoid in the circuit, and
measure voltage across the solenoid.
`
` 10 ohm _
` ___ / \
` .-|___|-----( A )-----.
` | \_/ |
` 10V | |
` --- .-.
` - Rx | |
` | | |
` | '-'
` | |
` '---------------------'
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

Even though this looks a little cheesy, it is in fact a true Kelvin
connection, unlike your Wheatstone setup. This measurement will be as
accurate as your ammeter. You don't need exactly 1 amp of current,
because you can just do the math using Ohms Law to get the inferred
resistance value:

R = V / I

Measure current each time you apply power -- it will change a little
for every time you get a different Rx. Try to keep the application of
power to the solenoid down to a few seconds or so if you can. That
will reduce self heating, which is always a problem in measuring
resistance of copper wire. Also look to ambient temperature, and prior
heating of the solenoid coil from use. Make sure the coil is cool
before you measure it.

Good luck
Chris

 

[Chris]

I guess I should mention one or two other things here. The Kelvin
measurement of resistance shown above didn't show the voltmeter:

` ___ / \
` .-|___|-( A )----.
` | 10 ohm \_/ |
` | o<----.
` | | |
` 10V | .-. / \
` --- Rx | | ( V )
` - | | \_/
` | '-' |
` | | |
` | o<----'
` | |
` '----------------'
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

Make sure you place the voltmeter leads right on the soleniod leads
where you want to make the measurement. If you place the leads on the
ammeter and the battery, you will be measuring the resistance of the
leads between the ammeter and the solenoid, and between the solenoid
and the battery. And also, you'll need a 10 ohm, _10 watt_ resistor.

10V * 1A = 10 watts, P = V * I

Second, a lot of people have those cheapie DVMs that only measure DC
current to 200mA or so. If you can't find a DVM that can measure DC
current of 1A, use a 100 ohm 1 watt resistor in place of the 10 ohm
resistor, giving you 100mA test current. You will only be measuring
20mV for an 0.2 ohm test resistor, but you'll still be way ahead of the
game if you've got a 200mV DC range on your DVM, despite reduced
voltmeter accuracy.

By the way, don't be too surprised if your calculated ohms are quite a
bit different than what the Wheatstone bridge led you to expect.
You'll be getting a more accurate reading this way, though.

If you have any questions or additional problems, don't hesitate to
post back.

Good luck
Chris

 

[Chris]

Matty F wrote:

Chris wrote:

Hi, Matty. You haven't said specifically what instrument you're using,
so I can't give specific advice. But I can say for sure that
Wheatstone bridges which only have two terminals for Rx aren't made to
measure low resistance. You might be getting self-heating in one of
the bridge's internal resistors, or if there's an active null, battery
voltage may be affecting that. These are the first causes that come to
mind.

I'll have a look at the instrument tomorrow. It's a quality instrument,
and I used an identical one 40 years ago at University. There are only
two terminals. I do suspect self-heating in an internal resistor.
I am being criticized for blaming the meter. Do you have any references
for inaccuracy in measuring low resistance, i.e. around 0.2 ohm?

Why don't you try using a 10 volt or so regulated power supply capable
of cranking an amp or so, a 10 ohm or so resistor, and a voltmeter and
ammeter. Connect the resistor in series with the power supply to give
you an approximately 1 amp source. Put the ammeter in series with the
circuit to measure current. Then put the solenoid in the circuit, and
measure voltage across the solenoid.

The boss has already suggested doing something like that, and is going
to bring his Avometer and other meters tomorrow. We have already put 30
amps through another solenoid, and it seems to work fine. We have a
large battery with a 100 amp meter and can change the current with a
compressible pile of carbon slabs.
All we want to do is to check whether all the solenoids are the same,
and whether they have shorted turns. Measuring the precise resistance is
not important. I was just wondering why it keeps changing.

In use, the solenoids get a variable high current (maybe 30 or more
amps) for a maximum of a minute or so.
According to tables I've just read, 100 feet of 14 AWG is about 0.25
ohms, which is in line with what I get.

Hi, Matty. The quality of the instrument doesn't matter if it's being
used to measure something it's not made to do. I'd be fairly confident
that if there are only two terminals, it's probably not made to
accurately measure less than an ohm. So don't worry about blaming the
instrument. From what you've said, that's the first place I'd look,
too.

Unless you have a resistance standard, you'll kind of have to wing it
on bouncing low ohm measurements. The first quick check is actually
the force current/measure voltage method. If you have a good handheld
DC voltmeter like the Fluke 77 (around 0.1% DC Volts accuracy), your
resistance measurement accuracy will be almost entirely dependent on
the accuracy of your ammeter. If you can crank 1 amp through your Rx,
another Fluke 77 measuring DC Amps will give you a resistance
measurement accuracy of 1.5%, which should be fine for what you're
doing. Inferring resistance by forcing a measured current and then
measuring voltage is theoretically sound as well as practical. Your
boss is making sense on this one. But in order to get more accuracy or
give you confidence in your measurements, Ohmite, Dale and others make
those aluminum-housed 50 watt 0.1% 100 milliohm Evanohm wirewound
resistors, which can be convenient, especially if you just want a quick
sanity check. If you thermal cycle the resistors in an oven several
times to stabilize them, measure resistance with a quality calibrated
instrument, and then make sure never to apply more than 1 amp of test
current to them (<20% of rated wattage), you can be fairly confident of
the results when you bounce instruments. I've got some myself in
decade values from .1 ohm to 100K ohm, and keep them in reserve for
when I need an ohms sanity check. All unofficial, of course.

High currents are more difficult to measure accurately, though. A
12VDC 1 amp unregulated wall wart, an LM317 with a heat sink, two
resistors and two 10uF caps will get you the regulated 10 volts. Most
of this stuff might even be in your junkbox or scrounge-able. Excess
currents will just cause Rx heating, and make your life more difficult.
You don't need and shouldn't use 30 or 100 amps to measure an 0.25 ohm
copper resistor.

Checking solenoids by checking resistance alone can be a bit of a
problem. The tightness of winding causes variations in wire length
which might account for your variations in resistance. Variations in
drawn wire diameter can make resistance values easily vary by 10% or
more. I think resistance _and_ inductance measurements might be a
better way of getting where you want. A better wound coil would have a
somewhat lower resistance, but a somewhat higher inductance. A coil
with marginally smaller wire diameter would have higher resistance, but
almost equal inductance. Missing or shorted windings would be more
noticeable with both measurements. If you've got one of those handheld
meters that measure inductance, that might help a lot.

Another practical, valid method you might use if you're serious is
putting a standard voltage across the solenoid, and measuring pull
force. I'm not sure if that's practical, though, for a 4-off check.

Good luck. Feel free to post again to let us know how you're doing.
Chris

 

[Chris]

Matty F wrote:

Chris wrote:

Hi, Matty. The quality of the instrument doesn't matter if it's being
used to measure something it's not made to do. I'd be fairly confident
that if there are only two terminals, it's probably not made to
accurately measure less than an ohm. So don't worry about blaming the
instrument. From what you've said, that's the first place I'd look,
too.

For the record I have found the following references saying that the
simple Wheatstone Bridge is unsuitable for measuring resistances under 1
ohm:

http://www.coe.uncc.edu/~elsheppa/2155/4-Wheatstone.pdf

Two of the more common types of bridges are the Wheatstone Bridge, which
is used to measure resistances of 1 ohm to 100,000 ohms, and the Kelvin
Double-Bridge, which is used to measure resistances in the range of
.0001 ohms to 1 ohm.

http://www.tpub.com/content/neets/14193/css/14193_94.htm

A Kelvin bridge is recommended for measuring resistances lower than 1 ohm.

http://www.8886.co.uk/forum/viewtopic.php?t=3

There are lots of "Wheatstone Bridge" circuits and the simplest of these
are NOT suitable for measuring low resistances below about 1 Ohm. So for
a typical resistance wire investigation the simple Wheatstone Bridge can
not be used.

http://operations.usace.army.mil/hydro/pdfs/bmp-Transformers.pdf

Test: Winding DC Resistance Measurements
Detects: Broken strands, loose connections, bad tap changer contacts.
Tool: Wheatstone Bridge (1 ohm and greater), Kelvin Bridge
micro-ohmmeter (Less than 1 ohm)

High currents are more difficult to measure accurately, though. A
12VDC 1 amp unregulated wall wart, an LM317 with a heat sink, two
resistors and two 10uF caps will get you the regulated 10 volts. Most
of this stuff might even be in your junkbox or scrounge-able.

I have analogue meters up to 100 amps that seem to work fine, and as
much DC power as anyone could want. There might even be a Kelvin Bridge
around here somewhere. Clearly we need something like that. I am only
trying to compare all the solenoids. I've already measured the pull
force for one solenoid and that's OK.
There's more than a dozen to check.

Hi, Matty. A 100 amp analog meter certainly isn't going to be good
enough here. Their accuracy is specified as a percentage of full
scale. Even a very good 100 amp meter might be specified as +/-2%,
which would be +/- 2 amps. You're not going to get there from where
you are. That meter will make you want to push higher current to get a
more accurate reading, which will cause Rx heating and ruin your
measurement.

The same is true for DC current applied. Don't think large -- you'll
ruin your measurement. Think small current, like 1 amp.

Get yourself a couple of cheapie $10 USD DVMs from the hardware store
if you have to. They'll be far more accurate for what you're doing
than a 100 amp analog meter.

And don't fall into the trap of thinking that unless you're reading
ohms directly off a brand name meter, the measurement can't be as good.
If you can accurately measure an appropriate DC current being forced
through Rx, and you can accurately measure the volts being impressed
across Rx, you _have_ measured Rx. It's ohm's law.

R = V / I

It works. The rest is just using common sense to avoid changing the
resistance too much with your measurement.

Good luck
Chris

 

[Chris]

Matty F wrote:

I have fed current to the solenoid and measured the current and voltage
with Avometers. The current was 1.45 amps (on the 1.5 amp scale) and the
voltage 0.38 volts, which works out to 0.26 ohms, which is the same as
what the Wheatstone bridge measured initially yesterday, and again
today. I've avoided pressing the test button too much this time.

The Wheatstone brand is "Pontavi" or some name like that. The name has a
strange font so I can't be certain it's right. I cannot see that name on
the Net. I'm told it has a 9 volt battery in it. If the lowest range
used two 5 ohm resistors in the bridge, when measuring a 0.25 ohm
resistance it would be trying to push nearly 2 amps through the bridge
resistors, which could heat them up to have the effect I got before when
the bridge readings kept going down for each measurement.
There is no "sensitivity" switch (which is badly needed).
There is no switch to put power on the resistors before the meter is
used, so the large impedance of the solenoid makes the meter move when
the test button is pressed.

Everybody but me in the workshop says that the Wheatstone bridge is fine
to measure resistances of way under 1 ohm. So clearly you and I and all
the references that I gave before are wrong. All my University work in
physics measurements and statistics was clearly wasted.
Anyway, thanks for your help.

Hi, Matty. Electronics is a specialized branch of physics. The trick
with the resistance is just a trick. You're on the right track here.

Two of the best electrical engineers I ever worked with started out
with degrees in physics.

You might want to check around at hamfests and such, and see if you can
find the manual for your "Punt-avi" bridge. It will tell you what you
need your coworkers to know.

Age doesn't necessarily bring wisdom, but it does bring experience.
Experience has taught me to RTFM (read the manual). If they don't have
the manual, get it.

It's funny sometimes (or so the War Department tells me) how much the
wisdom of age actually looks like just being tired. ;-)

Good luck
Chris