Maker Pro
Maker Pro

Sensing the flow of current

D

Daniel Rudy

Jan 1, 1970
0
Hello Everyone!

I've ran into a situation where I need to sense a very large flow of
current (100A) at 24VDC or so. This is for a motor supply feed. The
problem that I'm running into is isolation because there is a
microcontroller that is part of the control circuit that electrically
needs to stay away from this line as its powered from a different
source. So far, there are two ways that I'm aware of to do this:

1. Use a very low value resistor (I would need a few in parallel),
sense and amplify the voltage across that. Then feed that into a
voltage to frequency converter so it can be sent across an opto-isolator
and then on to a timer on the microcontroller.

2. Wind a couple of turns of the main feed wire onto a torridal core
form. The torrid has a gap cut into it. Inside that gap is a hall
effect device that is epoxied into place. This inherently provides the
galvanic isolation that is required. After signal conditioning, the
analog voltage can be sent directly to a ADC input on the controller.

I would prefer to use #2 as that would be easier to implement, and it
seems to have a lower parts count. Accuracy in the area of 1 amp
resolution or better, if possible, would be good. Any advise, links,
etc. is appreciated. Thanks.
 
G

Gary Reichlinger

Jan 1, 1970
0
I've ran into a situation where I need to sense a very large flow of
current (100A) at 24VDC or so. This is for a motor supply feed. The
problem that I'm running into is isolation because there is a
microcontroller that is part of the control circuit that electrically
needs to stay away from this line as its powered from a different
source.

It would seem that a 24VDC power supply and a microcontroller
power supply could be configured so that the grounds are connected.
If this were possible, then you could just put your resistors between
the low side of the power circuit and ground and then connect the high
side of the resistors to the ADC on the microcontroller. Even 8 bits
should give you the resolution you are asking for. If the isolation
is definitely required, there are other (more expensive) solutions.


So far, there are two ways that I'm aware of to do this:
1. Use a very low value resistor (I would need a few in parallel),
sense and amplify the voltage across that. Then feed that into a
voltage to frequency converter so it can be sent across an opto-isolator
and then on to a timer on the microcontroller.

International Rectifier has some prepacked chips that can do
this. They are generally used for higher voltage systems, but should
give you the accuracy you specify.
2. Wind a couple of turns of the main feed wire onto a torridal core
form. The torrid has a gap cut into it. Inside that gap is a hall
effect device that is epoxied into place. This inherently provides the
galvanic isolation that is required. After signal conditioning, the
analog voltage can be sent directly to a ADC input on the controller.

Allegro Microsystems has prepackaged sensors which operate on
this principal. (I am not sure if they go up to 100 amps.) Another
option is to use a magnetoresistive current sensor. F W Bell and
others offer these units.
 
S

scada

Jan 1, 1970
0
"Daniel Rudy"
Hello Everyone!

I've ran into a situation where I need to sense a very large flow of
current (100A) at 24VDC or so. This is for a motor supply feed. The
problem that I'm running into is isolation because there is a
microcontroller that is part of the control circuit that electrically
needs to stay away from this line as its powered from a different
source. So far, there are two ways that I'm aware of to do this:

1. Use a very low value resistor (I would need a few in parallel),
sense and amplify the voltage across that. Then feed that into a
voltage to frequency converter so it can be sent across an opto-isolator
and then on to a timer on the microcontroller.

2. Wind a couple of turns of the main feed wire onto a torridal core
form. The torrid has a gap cut into it. Inside that gap is a hall
effect device that is epoxied into place. This inherently provides the
galvanic isolation that is required. After signal conditioning, the
analog voltage can be sent directly to a ADC input on the controller.

I would prefer to use #2 as that would be easier to implement, and it
seems to have a lower parts count. Accuracy in the area of 1 amp
resolution or better, if possible, would be good. Any advise, links,
etc. is appreciated. Thanks.

--
Daniel Rudy

Email address has been encoded to reduce spam.
Remove all numbers, then remove invalid, email, no, and spam to reply.

http://www.allegromicro.com/sf/0750/ Has hall device ICs that are good for
100A. I just received some samples, but have not yet designed with them.
They look promising, high isolation and 0-5V output.
 
C

CFoley1064

Jan 1, 1970
0
Subject: Sensing the flow of current
From: Daniel Rudy
i0n1v2a3l4i5d6d7c8r9u0d1y2e3m4a5i6l7@n0o1p2a3c4b5e6l7l8s9p0a1m2.3n4e5t6
Date: 8/1/2004 7:42 AM Central Daylight Time
Message-id: <[email protected]>


Hello Everyone!

I've ran into a situation where I need to sense a very large flow of
current (100A) at 24VDC or so. This is for a motor supply feed. The
problem that I'm running into is isolation because there is a
microcontroller that is part of the control circuit that electrically
needs to stay away from this line as its powered from a different
source. So far, there are two ways that I'm aware of to do this:

1. Use a very low value resistor (I would need a few in parallel),
sense and amplify the voltage across that. Then feed that into a
voltage to frequency converter so it can be sent across an opto-isolator
and then on to a timer on the microcontroller.

2. Wind a couple of turns of the main feed wire onto a torridal core
form. The torrid has a gap cut into it. Inside that gap is a hall
effect device that is epoxied into place. This inherently provides the
galvanic isolation that is required. After signal conditioning, the
analog voltage can be sent directly to a ADC input on the controller.

I would prefer to use #2 as that would be easier to implement, and it
seems to have a lower parts count. Accuracy in the area of 1 amp
resolution or better, if possible, would be good. Any advise, links,
etc. is appreciated. Thanks.

Good morning, Dan. If you need isolated sensing of AC current, you would use a
current transformer. For DC current, the best way might be to use a hall
effect sensor. Several manufacturers make hall effect sensors built onto a
zero ohm shunt with integrated electronics. The current is then sensed by the
electrically isolated circuit, which gives you an analog output voltage
proportional to the DC current.

Allegro makes integrated current sensors which are good to up to 100 amps,
available off the shelf. One product you might want to look at is the
ACS750SCA-100, which operates on your microcontroller 5V supply. These are
available from stock at newark.com and other sources.

http://www.allegromicro.com/sf/0750/

Good luck
Chris
 
P

Peter A Forbes

Jan 1, 1970
0
Hello Everyone!

I've ran into a situation where I need to sense a very large flow of
current (100A) at 24VDC or so. This is for a motor supply feed. The
problem that I'm running into is isolation because there is a
microcontroller that is part of the control circuit that electrically
needs to stay away from this line as its powered from a different
source. So far, there are two ways that I'm aware of to do this:

1. Use a very low value resistor (I would need a few in parallel),
sense and amplify the voltage across that. Then feed that into a
voltage to frequency converter so it can be sent across an opto-isolator
and then on to a timer on the microcontroller.

2. Wind a couple of turns of the main feed wire onto a torridal core
form. The torrid has a gap cut into it. Inside that gap is a hall
effect device that is epoxied into place. This inherently provides the
galvanic isolation that is required. After signal conditioning, the
analog voltage can be sent directly to a ADC input on the controller.

I would prefer to use #2 as that would be easier to implement, and it
seems to have a lower parts count. Accuracy in the area of 1 amp
resolution or better, if possible, would be good. Any advise, links,
etc. is appreciated. Thanks.

I think that you can buy these sensors ready-made from people like LEME who use
them in their DC clamp meters.

We had a similar problem with some data loggers, but were able to use an RS 232
Galvanic Isolator between the loggers and the PC. The loggers were battery
powered. We had problems with currents flowing down the RS232 cable between the
PC ground and the equipment ground.

Peter
 
D

Daniel Rudy

Jan 1, 1970
0
And somewhere around the time of 08/01/2004 08:11, the world stopped and
listened as scada contributed the following to humanity:
http://www.allegromicro.com/sf/0750/ Has hall device ICs that are good for
100A. I just received some samples, but have not yet designed with them.
They look promising, high isolation and 0-5V output.

I went looking and found the ACS750ECA-100. Arrow has 6 peices
available for $8.43 each.
 
D

Daniel Rudy

Jan 1, 1970
0
And somewhere around the time of 08/01/2004 08:04, the world stopped and
listened as Gary Reichlinger contributed the following to humanity:
It would seem that a 24VDC power supply and a microcontroller
power supply could be configured so that the grounds are connected.
If this were possible, then you could just put your resistors between
the low side of the power circuit and ground and then connect the high
side of the resistors to the ADC on the microcontroller. Even 8 bits
should give you the resolution you are asking for. If the isolation
is definitely required, there are other (more expensive) solutions.

The problem has to do with the surge current. There can be a 70A or
higher current pulse on the surge when the motor kicks on under load.
Furthermore, that much current being suddenly drawn causes inductive
voltabe spikes along the wiring/pcb traces of the motor power circuit.
It is these spikes that I'm trying to keep away from the controller.
Allegro Microsystems has prepackaged sensors which operate on
this principal. (I am not sure if they go up to 100 amps.) Another
option is to use a magnetoresistive current sensor. F W Bell and
others offer these units.

I found some, they do.
 
J

Jamie

Jan 1, 1970
0
for DC current and good isolation we used
a Shut going through a torrid, the small
windings around the toride get a heavy
change effects in its inductance as the
current in the field causes it to collapse.
thus a DC current trough a torride can cause
the inductance to change. use this as part
of a Freq to Voltage converter circuit..
we also use a small mini circuit using a
shunt bar where both sides are input to the
+ and - side via resistors to an op-amp
using a dual supply., this supply is fully
ground isolated via a transformer type on the
secondary side so that the op-amp components do
not come in contact with any other electrical which
could potentially over voltage the op amp./
in this case the op-amp output drives an OP-coupler
with transistor output.
if you really want to get crude, you can use a
low voltage incandescent lamp accrossed the shunt in a
mini tube with a photo detector on the other end of the
tube.
 
G

Gary Reichlinger

Jan 1, 1970
0
And somewhere around the time of 08/01/2004 08:04, the world stopped and
listened as Gary Reichlinger contributed the following to humanity:

You posed the question. I just answered it. I never made any
claims as to the worldly importance of my comments.
The problem has to do with the surge current. There can be a 70A or
higher current pulse on the surge when the motor kicks on under load.
Furthermore, that much current being suddenly drawn causes inductive
voltabe spikes along the wiring/pcb traces of the motor power circuit.
It is these spikes that I'm trying to keep away from the controller.

There are ways of offering protection to the microcontroller
input that are much simpler and cheaper than full isolation. Usually
the ADC input will have a protection diode built in so that just
adding a resistor in series between the input and the test circuit
will be sufficient (see the microcontroller datasheet). Zener diodes
or a wide range of filtering circuits can also be used to protect the
input. It would seem unlikely that full isolation should be necessary
for a 24vdc system, but more details would be required to say that for
sure.
 
T

Tim Perry

Jan 1, 1970
0
"Daniel Rudy"
Hello Everyone!

I've ran into a situation where I need to sense a very large flow of
current (100A) at 24VDC or so. This is for a motor supply feed. The
problem that I'm running into is isolation because there is a
microcontroller that is part of the control circuit that electrically
needs to stay away from this line as its powered from a different
source. So far, there are two ways that I'm aware of to do this:

1. Use a very low value resistor (I would need a few in parallel),
sense and amplify the voltage across that. Then feed that into a
voltage to frequency converter so it can be sent across an opto-isolator
and then on to a timer on the microcontroller.

2. Wind a couple of turns of the main feed wire onto a torridal core
form. The torrid has a gap cut into it. Inside that gap is a hall
effect device that is epoxied into place. This inherently provides the
galvanic isolation that is required. After signal conditioning, the
analog voltage can be sent directly to a ADC input on the controller.

I would prefer to use #2 as that would be easier to implement, and it
seems to have a lower parts count. Accuracy in the area of 1 amp
resolution or better, if possible, would be good. Any advise, links,
etc. is appreciated. Thanks.

--
Daniel Rudy

Email address has been encoded to reduce spam.
Remove all numbers, then remove invalid, email, no, and spam to reply.

assuming this is a power supply and not a 24V battery, and assuming it is
unregulated: why not just put a standard toroidal pickup on the AC input to
the transformer? the resulting output it then rectified, filtered, and
calibrated.
any quiescent current will be so small in relation to the 100A load as to
be insignificant.

of course it might be a 3 phase input which would add to the complexity.
 
T

TP

Jan 1, 1970
0
Hello Everyone!

I've ran into a situation where I need to sense a very large flow of
current (100A) at 24VDC or so. This is for a motor supply feed. The
problem that I'm running into is isolation because there is a
microcontroller that is part of the control circuit that electrically
needs to stay away from this line as its powered from a different
source. So far, there are two ways that I'm aware of to do this:

1. Use a very low value resistor (I would need a few in parallel),
sense and amplify the voltage across that. Then feed that into a
voltage to frequency converter so it can be sent across an opto-isolator
and then on to a timer on the microcontroller.

2. Wind a couple of turns of the main feed wire onto a torridal core
form. The torrid has a gap cut into it. Inside that gap is a hall
effect device that is epoxied into place. This inherently provides the
galvanic isolation that is required. After signal conditioning, the
analog voltage can be sent directly to a ADC input on the controller.

I would prefer to use #2 as that would be easier to implement, and it
seems to have a lower parts count. Accuracy in the area of 1 amp
resolution or better, if possible, would be good. Any advise, links,
etc. is appreciated. Thanks.

Allegro Microsystems makes a 100A hall effect sensor ACS752xCA-100.
The two heavy leads connect inline with your motor current.
You apply 5 volts control supply. The zero-current signal is 2.5
volts. The signal
voltage increases for positive current and decreases for negative
current. You can send that to an ADC.
You can get a sample of the hall effect sensor from Allegro..

http://www.allegromicro.com/sf/0752/

TP
 
W

Watson A.Name - \Watt Sun, the Dark Remover\

Jan 1, 1970
0
Gary Reichlinger said:
On Sun, 01 Aug 2004 12:42:46 GMT, Daniel Rudy
<i0n1v2a3l4i5d6d7c8r9u0d1y2e3m4a5i6l7@n0o1p2a3c4b5e6l7l8s9p0a1m2.3n4e5t6

It would seem that a 24VDC power supply and a microcontroller
power supply could be configured so that the grounds are connected.
If this were possible, then you could just put your resistors between
the low side of the power circuit and ground and then connect the high
side of the resistors to the ADC on the microcontroller. Even 8 bits
should give you the resolution you are asking for. If the isolation
is definitely required, there are other (more expensive) solutions.


So far, there are two ways that I'm aware of to do this:

International Rectifier has some prepacked chips that can do
this. They are generally used for higher voltage systems, but should
give you the accuracy you specify.

Allegro Microsystems has prepackaged sensors which operate on
this principal. (I am not sure if they go up to 100 amps.) Another
option is to use a magnetoresistive current sensor. F W Bell and
others offer these units.

Third way. Run the current thru a core of an inductor, which has the
winding for an oscillator. The DC varies the coil's inductance which
changes the freq of oscillation. Isolation is inherent as in the above.
 
J

Jim Adney

Jan 1, 1970
0
2. Wind a couple of turns of the main feed wire onto a torridal core
form. The torrid has a gap cut into it. Inside that gap is a hall
effect device that is epoxied into place. This inherently provides the
galvanic isolation that is required. After signal conditioning, the
analog voltage can be sent directly to a ADC input on the controller.

These are available as a commercial product from a Swiss company
called LEM. They make a good product line.

-
 
E

Ed Price

Jan 1, 1970
0
It ought to be embarrassing for a poster to ask a serious question, have it
seriously answered, and then reply with an automatic tag file that's
sarcastic and demeaning.

But I suppose Mr. Rudy is either too dumb to recognize his automated insult,
or too sophomoric to care.


Ed
wb6wsn
 
D

Daniel Rudy

Jan 1, 1970
0
At about the time of 08/01/2004 11:37, Gary Reichlinger stated the
following:
You posed the question. I just answered it. I never made any
claims as to the worldly importance of my comments.

Sorry about that, no insult was implied or intended. It's one of those
automated taglines I came up with a long time ago. Just haven't gotten
around to changing it and completely forgot about it. It's changed now.
There are ways of offering protection to the microcontroller
input that are much simpler and cheaper than full isolation. Usually
the ADC input will have a protection diode built in so that just
adding a resistor in series between the input and the test circuit
will be sufficient (see the microcontroller datasheet). Zener diodes
or a wide range of filtering circuits can also be used to protect the
input. It would seem unlikely that full isolation should be necessary
for a 24vdc system, but more details would be required to say that for
sure.

The other reason that I'm going for full isolation is that the
microcontroller and motor are powered from completely separate supplies.
 
R

Robert Baer

Jan 1, 1970
0
Daniel said:
Hello Everyone!

I've ran into a situation where I need to sense a very large flow of
current (100A) at 24VDC or so. This is for a motor supply feed. The
problem that I'm running into is isolation because there is a
microcontroller that is part of the control circuit that electrically
needs to stay away from this line as its powered from a different
source. So far, there are two ways that I'm aware of to do this:

1. Use a very low value resistor (I would need a few in parallel),
sense and amplify the voltage across that. Then feed that into a
voltage to frequency converter so it can be sent across an opto-isolator
and then on to a timer on the microcontroller.

2. Wind a couple of turns of the main feed wire onto a torridal core
form. The torrid has a gap cut into it. Inside that gap is a hall
effect device that is epoxied into place. This inherently provides the
galvanic isolation that is required. After signal conditioning, the
analog voltage can be sent directly to a ADC input on the controller.

I would prefer to use #2 as that would be easier to implement, and it
seems to have a lower parts count. Accuracy in the area of 1 amp
resolution or better, if possible, would be good. Any advise, links,
etc. is appreciated. Thanks.

--
Daniel Rudy

Email address has been encoded to reduce spam.
Remove all numbers, then remove invalid, email, no, and spam to reply.

Since you state DC current, and high voltage isolation, ther are two
solution types:
1) Hall effect transducer. These can do the job, but one cannot get 1%
accuracy, linearity or temperature insensitivity that you seem to
demand.
But if you are willing to put up with about 12% nonlinearity with
isolationof 5nA at 2000V, look at the Zetex ZMC20 (need to modify
magnetic coupling to allow 100A sensing).
2) Current transformer (fluxgate) as made by LEM. Expensive, but gives
wideband and precision response. Perhaps you could "roll your own".
 
J

John Doe

Jan 1, 1970
0
Hi Daniel,

Another way: If you know the resistance/lenght of the wire, you may put
another big resinstance in paralell and sense the current trough the
later. I1=R2·I2/R1 where R2 is the small wire resistance, and R1 is the
big paralell one. Anyhow, I recommend to isolate from there before
entering the microcontroller.
 
Top