I can’t find it now but I saw a you-tube video where someone used a miliamp meter and some kind of homemade pickup to display amps. He was reading miliamps and making the conversion to amps by moving the decimal point. It wasn’t the focus of the video so I can’t look it up. Anyone know how to build this? I tried Googling but had no luck.
Sounds like something along the lines of a current transformer (CT). A 1000:1 CT will output a current of 1 mA when 1 A is passed through it. We use lots of CTs for branch and subfeed distribution monitoring in our products, of various ratios, such as 250:1, 600:1, 600:5, etc.
Easy, but calibration can be tricky.
Get yourself whatever remote meter you want. I recommend a volt meter, not an Amp meter (you’ll see why in a second). Then, take a length of heavy copper wire (how heavy is dependent on how much current you want to measure), and make a basic calculation of how long you will need, using a chart like this:
This will become your “shunt.”
So, if you want to measure 20 Amps, and have it register 20 mV on the meter, you need to have .001Ω of wire, which is about 1 foot of 10 Gauge wire. Then you cut a little extra, and solder one lead from the meter close to the end of the wire, and leave the other free. Then, you can pass a know current through the wire, and slide the free end STARTING AT THE SOLDERED END down the length of the shunt, until the meter reads correctly. Solder the wire at that point, and carefully coil the shunt, and you’re are done.
You can also buy pre-made shunts online (digikey, for example).
On edit: QEDs current transformers are a good solution, but only for AC. They are also inaccurate if run long distances.
Likewise, yours is a good solution for either, but your accuracy will suffer with increasing current, since heating of your shunt resistance will (usually, depending on the materials used) cause the resistance to increase. There are various schemes you can use to compensate, but all of them naturally increase the complexity of the design. Of course, how much accuracy you actually need depends greatly on your application.
I love the internet. Kids have no clue what a gift it is.
OK, I printed out the ePanorama.net info. Not quite sure what I’m looking at. I want to be able to measure amps on a 12 V DC system between 0 and 100 amps. I would like to be within 1 amp of accuracy. Also not sure how to run a known amount of amps as a test. I looked up shunts on Wiki and it looks like a resistor dropped in series. Is that something that could be used for ongoing measurements? It looks like it would limit current flow.
All current measurement systems introduce some error. Most high-current shunts are extremely low resistance, so the amount of voltage drop is (or should be) insignificant. A .001Ω current shunt will introduce .1V of drop at 100A, so your 12 volts will then be 11.9 volts, and your meter would measure 100mV full scale and 1mV at 1A. If you have an even more sensitive meter, you can get away with an even lower resistance shunt.
in the tradition of stupid-human tricks I just drove a pair if mini dykes into my thumb trying to strip wire. Blood everywhere. My house looks like a crime scene.
I think I understand this now. I originally thought the wire was used as an inductive coil like a clamp meter. What you describe is using the natural resistance found in wire. The current to be measured is going THROUGH the wire. Does the resistence change with the number of coils or is that done to make it shorter.
Sorry about the thumb. (You’ve just been self-initiated into the IEEE).
I guess I should have made it clearer - yes, the only reason to coil the wire is to make it more compact. A commercial shunt will use a huge chunk of metal, but that’s not easy to work with as a DIYer - so you need to use a thinner, longer piece of wire, and coiling it just gets it into a smaller space.
We have quite a few current shunts in our metrology lab. They’re tricky to use. You have to be careful hooking them up to instrumentation with non-isolated inputs. You have to use a temperature calibration curve. And drift due to thermoelectric effects will cause all kinds of problems. (“T/C offset compensation” is sometimes necessary.) Properly used, however, they are extremely accurate.
Because of the problems associated with them, current shunts are not used a whole lot in industry nowadays. For most industrial applications, the best approach is to use an AC + DC current transducer. A series transformer is used to measure the AC component of the current, while a Hall Effect sensor is used to measure the DC component. Such a device has an isolated output, is fairly “non-intrusive,” and has a linear transfer function (V/A).
Yeah, but…
The OP wanted a DIY way to measure current to 1% accuracy. A copper shunt that is dropping 100mV at 100 A is only dissipating 10W. Over a 1’ piece of wire, the self-heating is going to be very small, and all the issues with temperature and junction effects are going to contribute less than 1% error.
I believe he said ±1 A, not ±1%. If he’s looking for ±1% error from 0 to 100 ADC, a simple wire won’t work. He will need to shell out some big bucks. Trust me on this.
If he needs ±1 A uncertainty from 0 to to 100 ADC, a current probe is the way to go. A couple that will work:
Extech 38394
http://www.transcat.com/PDF/38394.pdf
DC Current Range: 0 to 600 ADC
Price: Around $100
Amprobe ACDC-100
http://www.transcat.com/PDF/ACDC100.pdf
DC Current Range: 0 to 400 ADC
Price: Around $100
The Amprobe will be a bit more accurate than the Extech. Because of the wide range (0 to 400 ADC), relative accuracy will begin to suffer for very low currents. At 1 amp, for example, the error will be about ±0.4 A (which is ± 40%). The relative error is much less at higher currents. If the uncertainty requirement is ±1 A from 0 to 100 ADC, then this will work nicely.
The other idea, of course, is to use a shunt resistor or wire. If you do this, Magiver, stick the resistor on the low side, not the high side. Picking a value will be tricky. It will be a balancing act between trying to minimize heat dissipation & voltage drop at high currents vs. trying to get a good voltage reading at low currents. If you use a 0.001 Ω resistor, for example, you’ll get a fairly robust signal at 100 ADC (100 mV) and the heat dissipation won’t be too bad (10 W). But at 1 A the signal will only be 1 mV. Accurate measurements at this range requires careful techniques and a fair amount of money. To get around this problem, you could “switch in” different resistors depending on the current. But be careful… if you have the switch set incorrectly you could end up cooking some resistors.
If the uncertainty requirement is ±1 A from 0 to 100 ADC, a better solution is to use the Amprobe ACDC-100. 
Magiver: I work in Dayton. Give me a buzz: magcraft@windstream.net.
All good points. May I just add that the change in resistance with temperature is, at least, pretty predictable. For metallic elements in general, resistance is approximately proportional to absolute temperature. This isn’t true of alloys, especially for alloys like manganin or chromel or nichrome that were developed to have resistances that are independant of temperature. Fortunately, copper is unusually simple in its resistance as a function of temperature, being proportional to an accuracy unmatched by other common metals. Or, at least, being linear - I forget which - but I think it’s proportional.
Another option is to use multiple wires in parallel as your high power, low resistance shunt. You have to do the voltage measurements separately on each, though.
Yep. Back when I worked for a transformer manufacturing company, we used this principle to gauge the interior temperature rise of the windings.
The standard shunts we use in our calibration lab incorporate thermocouples or thermistors for measuring temperature. After measuring the temperature, the resistance deviation is calculated using a deviation curve supplied by the primary standards lab. The alloy is usually formulated such that the slope of the curve is around zero at room temperature. Same goes for the Thomas standard resistors we maintain in the oil bath.
Compensating for temperature is the easy part. Making the voltage measurement is the hard part. It’s not simply a matter of sticking a voltmeter across the shunt and making a measurement. Making accurate measurements down in the millivolt and microvolt range requires a *lot * of experience, expensive equipment, and a ton of patience.
The TCR of copper is really pretty bad - .4%/°C
Here is a ready-made solution that is undoubtedly more accurate.
There’s lots of high-current measurement stuff on ebay (mostly for the RE crowd).
:dubious:
Yeah, that’s debatable, but it’s got to be better than the 1% over 10° F of pure copper.
That’s the meter accuracy (probably at full scale) not the accuracy of the shunt, assuming I’m reading the poorly-worded ad corrrectly.
Doubtful, even if it’s a “full-scale” accuracy.
First off, it’s impossible for a digital meter to have an uncertainty of “0.1% of reading” over the entire range; there must also be an absolute uncertainty added to it. Secondly, a digital meter usually does not have an absolute uncertainty of ±1 count. It’s usually somewhere between ±2 and ±8 counts (though it could certainly be ±1 count for a meter that isn’t very sensitive).
If I were to make a WAG, I’d say the accuracy of that meter is ±(2% of reading + 0.5) amps. And that’s being generous.
Though it would be a little more expensive of an option, I’d use a clamp-on probe w/ Hall Effect sensor vs. the kludged-together meter from eBay. The former will be more reliable, more rugged, exhibit a much lower voltage & power burden, and will be capable of performing other measurement functions as well.
Wow! Thanks for the responses. I can’t get over the power of computers and the internet. When I was a kid the idea of getting on a world network and researching information was pure fantasy. The ability to ask an open question to the world and getting a response is so incredible.
I’m going to try calibrating my own shunt. It shouldn’t be too difficult to compare the numbers to a clamp meter, which is on my toy list.