Contactless DC current measurement

[ikalogic]

hello everybody,

At work i am asked to develop a mean of detection the flow of current in a conductor. The current is DC current, should be about 50mA.

I've tried many different configuration to build a "kind-of" current transformer, with 5 to 10 turns around the conductor under test. the two free leads of the pick-up coil are connected to a resistor, and fed to an op-amp..

I've tried many different configurations, but i don't seem to be getting satisfactory results.. no clear voltage change with respect current change. (i just want to detect a variation between 0 mA and 50 mA)..

So before i post a schematic of my trials and share it here.. can anybody experienced tell me if it is doable at all?

It should be since DC current clamps exists...

I'll keep on trying.. waiting for some response-advice from your side.. i am really lost on that project...

 

[Nigel Goodwin]

You're trying to build a transformer, a transformer only works on AC - I think the DC ranges on clamp meters work using direct connection and not the clamp.

You could perhaps try a hall effect sensor?, which can measure a steady magnetic field.

 

[ikalogic]

You're trying to build a transformer, a transformer only works on AC - I think the DC ranges on clamp meters work using direct connection and not the clamp.

You could perhaps try a hall effect sensor?, which can measure a steady magnetic field.

yes.. i am trying to think about that, but is to possible to "home-make" a cheap hall effect sensor? or the only option is to buy the ready one? because as far as i understand. it's simply a pickup-coil and an amplifier.. right?

Alternatively, would it be possible to detect a "peak" at the instant that the curents starts/stops flowing? (this, sure could be done using a 555 or any microcontroller, but my question is: is a single transition alternative enough to generate emf and be detected?)

 

[Pommie]

A hall effect sensor is not a pick up coil. A pick up coil can only ever measure AC. Hall effect sensors can measure DC.

To increase the sensitivity of a HE sensor you can place it in a ferrite ring like so,
**broken link removed**

This will make it much more sensitive. If you pass the conducting wire through the ring a number of times it will become even more sensitive.

Mike.

 

[crutschow]

yes.. i am trying to think about that, but is to possible to "home-make" a cheap hall effect sensor? or the only option is to buy the ready one? because as far as i understand. it's simply a pickup-coil and an amplifier.. right?

Alternatively, would it be possible to detect a "peak" at the instant that the curents starts/stops flowing? (this, sure could be done using a 555 or any microcontroller, but my question is: is a single transition alternative enough to generate emf and be detected?)

I believe some commercial dc current sensors apply a high frequency AC voltage to a coil wrapped around a small magnetic toroid. The dc current through the center of the toroid affects the core saturation which is reflected in a change in the AC current. But this is a technique that requires significant auxiliary circuits.

A purchased hall effect sensor is probably the easiest way to do a home brew dc current sensor. A hall effect sensor is not "a pick-up coil and an amplifier", it is a special semiconductor device that gives an output voltage proportional to a magnetic field through it. You can not make one at home any more than you could make a transistor.

You could detect the change is current as it starts flowing or changes value with an ac current transformer but, of course, that would give you no steady-state value. You need an amplifier to detect this change. A 555 is a timer which would be useless for this purpose and a microcontroller is a digital device (even if it has an A/D input, it would likely require amplification of the signal).

 

[ikalogic]

So as it seems, my best shot is Hall effect sensor.. all the rest is not quite realistic..right?

I even tried to detect that peak i was talking about.. but seems like 40mAmps wont generate that much current in the coil...

 

[ikalogic]

A hall effect sensor is not a pick up coil. A pick up coil can only ever measure AC. Hall effect sensors can measure DC.

To increase the sensitivity of a HE sensor you can place it in a ferrite ring like so,
**broken link removed**

This will make it much more sensitive. If you pass the conducting wire through the ring a number of times it will become even more sensitive.

Mike.

So, mike, i am going in your way: hall effect. However, i'll have to convince the rest of the team at work to use that sensor.. the problem is price!

I've found that cheap hall effect sensor, seems to me very adequate to my application "detection the flow of a DC current of 50 to 100 mA"
https://www.electro-tech-online.com/custompdfs/2008/06/3245.pdf

Can you or someone else please take a look at the datahseet and tell me if it seems adapted to my application.. ?

thanks a lot.. i think i'me going to fill an order for a couple of those to make some tests..

 

[Mikebits]

Check out this link.
**broken link removed**

 

[ikalogic]

Thanks..

Actually we are searching for a cost effective solution.. the sensor i proposed costs less than 30 cents..

 

[crutschow]

I've found that cheap hall effect sensor, seems to me very adequate to my application "detection the flow of a DC current of 50 to 100 mA"
https://www.electro-tech-online.com/custompdfs/2008/06/3245-1.pdf

Can you or someone else please take a look at the datahseet and tell me if it seems adapted to my application.. ?

The selected device has a digital (comparator) type output which is low with no field and high above the trigger point. Is that what you want?

If you just need to detect the presence of a current and not its analog value then another (and possibly cheaper) way would be a reed relay. You wrap a coil of wire around the relay with the current flow and its contacts will close when the current flow (and consequent magnetic field) are high enough. You adjust the number of turns to get the required sensitivity.

 

[Ubergeek63]

you either need to get a linear and fight with the idiosyncrasies or get a real hall current sensor, which at $3 in singles is less then the crap you are going to have to hack together for what you are proposing

 

[Mikebits]

Unless your building thousands, I fail to understand why cost is such an issue. What is that old saying? Something about a penny wise or something...

 

[Nigel Goodwin]

You can find hall-effect sensors in DC brushless motors (as used in VCR's, HDD's etc.), I've no idea if you could use one of those, but as you can find them for free, it could be worth a little experimenting?.

 

[Pommie]

A while ago I got one of these to have a play with. They cost a few dollars and are not very sensitive. I hooked it up to a pic chip and put a ferrite ring with a slot on it. The ring had 8 turns of wire on it and with no current I got a reading (from the ADC) of 514 - roughly half way = 2.5V. With 1 amp flowing I got a reading of 528, at 2A I got 542, so pretty linear. If you used thinner wire (I had thick PVC wire) and more turns you should get a reasonably accurate reading. As I'm writing this I just rewound it with 25 turns of thinner wire (0.6mm) and I now get a reading of 550 at 1A.

The hardest bit was cutting the slot in the ferrite ring. I used my dremel with a metal cutting disk and it was very slow and chipped lots.

Mike.

 

[ikalogic]

Unless your building thousands, I fail to understand why cost is such an issue. What is that old saying? Something about a penny wise or something...

The first production will be over 50 000 pieces, then, yearly production will be around 180 000 parts.

 

[ikalogic]

thanks guys,

I am on a completely different lead.. actually i am measuring the voltage drop on the battery, amplify the noise, and make an FFT on the amplified noise, and detect a recurrent voltage drop that happens always at the same frequency, with corresponds to my load which is triggered at 1.48 Hz.. seems to be working pretty well....

 

[crutschow]

Your pulling my leg, right? How can taking the FFT of a signal be a simple, low-cost way to measure the current?

 

[dknguyen]

Maybe with an oscilloscope whose cost is not being accounted for?

 

[crutschow]

Maybe with an oscilloscope whose cost is not being accounted for?

Well, he talked about a "low cost solution" and a yearly production of 180,000 parts so I just assumed (incorrectly apparently) that an expensive instrument that measures FFT was not in consideration as a possible solution. Heck, he can buy a DC current probe from Tektronix for a few thousand bucks (US) that'll do the job too.

Apparently this is for a production test and not some device that is built into the unit, but we we never told that. It's a waste of time for the rest of us if some ----- does not tell us what he really needs and we go off on a solution tangent.

 

[mneary]

I assumed he's going to take the noise on the supply (assuming that the supply impedance is predictable enough), amplify it to meet the uC's input requirements, and do a one-frequency software fft.