up to ~300A current sensor

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
User avatar
kheston
Posts: 348
Joined: Wed Dec 03, 2003 1:01 am
Location: CA
Contact:

up to ~300A current sensor

Post by kheston » Fri May 30, 2008 5:53 pm

I have an Allegro Micro ACS75 that gives me +/- 50A. I've mounted it on a copper shunt that splits the current well enough to give me the readings I need.

I don't like the solution very much, frankly. It's pretty fragile. I guess I can stiffen things up a bit and make it more durable but I'm hoping there's a way to leverage the Hall Effect to do something similar.

All of the DIY posts I've seen deal with much smaller currents.

I'm picturing a coil around a length of PVC that I poke my battery cable through. I don't need the resolution the ACS75 is capable of: nearest 5A would suffice.

Can anyone point me in the right direction?
Kurt - SF Bay

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Fri May 30, 2008 7:46 pm

Hi there Kurt,

From what i understand you want to measure current, and you
want to measure dc current and it's higher than 50 amps.
The device is rated for 50 amps so that means you have to
shunt some current around it, although you seem to have this
working already.
You also seem to want to use a coil, which i dont understand yet.
A coil would increase the magnetic field which would activate
the Hall sensor even more, which would make it read 50 amps when
you are really have only 25 amps. This seems to be contrary to
what you are looking for. This is when you run the current through
the coil, which is the only thing possible with dc.
Also, the coil wrapped around the plastic pipe wont do much good
when working with dc current.

What is the top end of the current you need to measure...
100 amps, 200 amps, 500 amps?

You know they make higher current models too right?
LEDs vs Bulbs, LEDs are winning.

Robert Reed
Posts: 2276
Joined: Wed Nov 24, 2004 1:01 am
Location: ASHTABULA,OHIO
Contact:

Post by Robert Reed » Fri May 30, 2008 7:53 pm

"I'm picturing a coil around a length of PVC that I poke my battery cable through. I don't need the resolution the ACS75 is capable of: nearest 5A would suffice."

I assume you are measuring DC currents by this statement. Your coil around PVC sounds like a current transformer for AC USE only. Have you considered making tour own current shunt? At 300 amps, a length of poor conductor material (as compared to copper) such as hot rolled steel will give you an out put that woud develop enough voltage (in the millivolt range) that you could work with for further amplification, scaling and calibration. At 300 amps,this would have to be a sturdy device with 1/4" bolt connections to the line, 24 ga. hookup wire from there to a high impedance device. Of course there would be a certain degree of experimentation to perfect. If you have alredy priced 300 amp shunts its worth a shot.

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Fri May 30, 2008 8:01 pm

Hi again,

Robert, you posted just minutes after mine :smile:
My previous post addressed some of the issues you were
talking about.

Just to note, i've made pretty good shunts out of brass too,
the strips you get at the hobbie shops. They make pretty
good shunts and you can file the center to calibrate.

Since he already has a nice Hall sensor, i would say use that.
If the shunt isnt mechanically stable, heck, some epoxy should
help things out. Enclose it into a box perhaps?
LEDs vs Bulbs, LEDs are winning.

User avatar
kheston
Posts: 348
Joined: Wed Dec 03, 2003 1:01 am
Location: CA
Contact:

Post by kheston » Fri May 30, 2008 8:10 pm

Al,

I was hoping for suggestions about how one might build his own high-current-to-low-voltage transducer. I was thinking (perhaps incorrectly) that if I don't need the 0.75 - 2% accuracy the Allegro stuff offers that I might be able to replace the device altogether with a DIY hall-effect sensor. My understanding is the 400A DC clamp meter I bought for $90 uses the hall effect in a similar fashion.

So, I guess I'm asking whether I'd be chasing rainbows with a split ferrite core and some copper wire in hand hoping to get a predictable voltage to measure.

Like I said, I already have something working that I can improve upon. Just wondering if the effort might be better spent building something custom from scratch. With all the machining time I'll need to put into my custom shunt, I just want to make sure the Allegro device is the best thing to design around.

Looks like Allegro will go up to 200A unidirectional. Since I needed a shunt anyway and wanted bi-directional, I bought the +/-50A one they had in stock.

BTW - I'm controlling my generator's output very well with the circuit you gave me. I have a range from about 2A to ~60A on each of the two alternators I've tested so far. Perfect! This thread is about getting the right data to adjust the digipot with to keep from cooking my battery bank.

Thanks for all the hand-holding!

--K
Kurt - SF Bay

User avatar
kheston
Posts: 348
Joined: Wed Dec 03, 2003 1:01 am
Location: CA
Contact:

Post by kheston » Fri May 30, 2008 8:21 pm

R^2,

I did look into the big DC current sensors at Digikey ($700+, ouch). Hard sell with the wife and no learning-by-doing value. Hence, my post.

Hot rolled steel, can get some of that...the reading I'm doing about op-amps may get me to where I can read those mv with an ADC, too. Got any links that may shorten my learning curve?

Thanks,
K
Kurt - SF Bay

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Fri May 30, 2008 10:20 pm

Hi again Kurt,

Well ok, then my next question would be what kind of temperature
range are you looking for, and also, do you really need isolation?

The Hall effect sensor basically gives you isolation even with DC
current measurement, something that is hard to get without one.
For AC of course there is always the current transformer, but
with DC the only option is Hall effect or special linear opto isolator.
This is why i ask about the isolation, because in many cases you
dont really need actual isolation so why bother (brings in added expense).
If you can do without this you can work up a construction that is
pretty cheap.

Also, since you already have a 400 amp meter, you can use that to
calibrate a moon rock to measure current with your cheap DVM :smile:
Ok, more seriously, you can build your own shunt and use an op
amp (with low offset voltage) to measure your DC current for your
feedback (presumably).
Another point when building your own amp/shunt is that you seldom
have to have it read to zero with any accuracy, because your circuit
is normally working at some mid to high range anyway, so you can
do a little testing in this area and perhaps a minimal calibration to
get things working to a very decent accuracy, probably more than you
need.

The tricks are fairly simple:
1. Find a low input offset op amp (much lower than 3mv as with the
common type. This usually requires something like OP07 or similar
(if i rem my part number correctly here).
2. Build a current shunt based on the resistance of your material,
such as copper or brass. You can look up the resistance of the metal
and calculate the resistance for a given length width and height of the
material you choose. Figure on a little temperature variation because
that's the way metals are. Many times it wont matter that much anyway
because the end product doesnt need super regulation anyway, just
something that gets you in the right general area. You do need heavy
construction here though, as Robert says quarter inch bolts would be
a good idea for the electrical connections.
3. Construct the circuit with the op amp and some decent gain, then
check for accuracy with the 400 amp meter. If you need to adjust one
of the gain resistors, you can do that to tune the accuracy in better.

That's about it. Not too complicated huh?

I've actually done this in the past. The current wasnt as high as yours
is though (around 20 amps max) but the principle is the same.
I didnt even use a low offset amp...what i did was adjusted the
zero offset of my analog meter instead to make up for the input offset
error in the cheap op amp (ha ha ha) but hey, it worked.
Used it for several years that way, no problems.

Oh yeah, you need a small power supply also to power the op amp,
so maybe a LM317 regulator or something like that. The nice
op amps need to get down to zero volts if your measurement is
that low too, so you might end up having to use a negative supply
too. If you have to work off of a battery that may mean building
a small dc to dc converter that provides a minus voltage from a
plus voltage, but there are a number of switched capacitor converters
out there that should provide the needed negative supply, if that
becomes necessary. They are not expensive either.

I guess the most important question that needs to be answered before
any real circuit can be considered is, does this app really need isolation
or not? If you want to have isolation using a Hall effect sensor, then
that's a different story. The linear Hall effect sensor puts out a voltage
that is related to the magnetic field, which in turn is related to the current
flowing in a nearby wire. The trick here is to get the Hall sensor
mounted next to the wire so that it can not move in any direction or
rotate relative to the wire. You can then measure the output voltage
and go from there.
LEDs vs Bulbs, LEDs are winning.

User avatar
Externet
Posts: 1862
Joined: Fri Feb 01, 2002 1:01 am
Location: Mideast USA
Contact:

Post by Externet » Fri May 30, 2008 10:27 pm

- Abolish the deciBel ! -

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Fri May 30, 2008 10:35 pm

Hi Miguel,

Oh nice link :smile:
I am tempted to buy one myself, for only 30 dollars each.
Or, i might build one with a 50 cent op amp and 90 cents
piece of brass and a few 1/4 watt resistors (hee hee).

Would be nice to have one already calibrated though, as im
sure those would be.
LEDs vs Bulbs, LEDs are winning.

User avatar
kheston
Posts: 348
Joined: Wed Dec 03, 2003 1:01 am
Location: CA
Contact:

Post by kheston » Fri May 30, 2008 10:52 pm

Al,

Thanks for the great post with all the detail. Very helpful.

I'm pretty sure I don't need any isolation, but if I run into weirdness I'll post (because I don't really know for sure).

I need to hit the metal yard for some other supplies so I'll be sure and pick up an appropriately sized hunk of brass or other suitable metal that I find on Google.

Thanks again.

--K
Kurt - SF Bay

dyarker
Posts: 1752
Joined: Fri Aug 22, 2003 1:01 am
Location: Izmir, Turkiye; from Rochester, NY
Contact:

Post by dyarker » Sat May 31, 2008 6:13 am

On the page Externet gave the link for, notice the small screw terminals on the side of some of the shunts? They are for the sense wires. That is so the op amp only measures the voltage across calibrated part of the shunt. If the sense wires are connected to the large terminals, you would also be measuring any (uncalibrated) voltage drop between the large terminal lugs and the body of the shunt. Even with clean, tight, connections, the error could be two digits of percent.

Suggest if you make your own shunts, that you also make separate termination points for the sense wires.

Cheers,
Dale Y

Robert Reed
Posts: 2276
Joined: Wed Nov 24, 2004 1:01 am
Location: ASHTABULA,OHIO
Contact:

Post by Robert Reed » Sat May 31, 2008 8:01 am

Only one thing to add to all the replys - if you do decide to make your own shunt, just make sure it is large enough to stay cool under load. Even a small amoumt of heating will change the shunts resistance and give errors in your readings.

User avatar
Bob Scott
Posts: 1192
Joined: Wed Nov 20, 2002 1:01 am
Location: Vancouver, BC
Contact:

Post by Bob Scott » Sat May 31, 2008 11:10 am

MrAl wrote:I am tempted to buy one myself, for only 30 dollars each.
Or, i might build one with a 50 cent op amp and 90 cents piece of brass and a few 1/4 watt resistors (hee hee).

Would be nice to have one already calibrated though, as im
sure those would be.
Once upon a time I made a 10 Amp shunt out of a piece of coat hanger wire. (Coat hangers and stainless steel car radio antennas are useful for so many mechanical quick fixes too. eg: small hooks, push-pull links, etc.) Yes, I had to calibrate it!

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Sat May 31, 2008 3:00 pm

Hi again,

Oh ok Bob, that sounds nice and cheap too.

Here is a design for an 80 amp current shunt, actually two of them.
It's going to take a little more brass than the 10 amp shunt however.

As Robert points out the shunt has to be big enough to handle the
required power dissipation at the rated current, and also that
the high current connections should be nice heavy bolts, and as dyarker
points out the sense terminal connections to the shunt should be
separate from the heavy current connections.

With all this in mind, here are two designs for an 80 amp shunt...


Image
LEDs vs Bulbs, LEDs are winning.

rshayes
Posts: 1286
Joined: Tue Mar 04, 2003 1:01 am
Contact:

Post by rshayes » Sat May 31, 2008 8:19 pm

The stacked design that MrAl shows can be improved by making the high current connection at the top of the stack on one side and at the bottom of the stack on the other. This makes the number of joints in each current path equal. The resistance in each contact may be significant compared to the resistance of the brass strip.

When connections are made at the top on each side, more currrent will flow in the top strip than in the other two. The distribution of current will depend on the contact resistances.

Connecting to the top on one side and the bottom on the other equalizes the currents through the three strips if the contact resistances are equal, and this will equalize the voltage drops across the three strips.

If the voltages on the three strips are equal, the measurement connections could be made to the top or bottom strip. This might be easier to construct.

The measurement connections could also be made by extending the center strip and moving the measurement connections to the outside of the spacers.

Post Reply

Who is online

Users browsing this forum: Google [Bot] and 45 guests