Measuring <20 uA currents in 1.55 volt circuit???

[handymandave]

Hello,
I am a watchmaker who works on quartz watches, among other things. (www.dave-arnold.com) I am not looking for work!
I need someone's expertise. I need to accurately measure low DC currents, mostly under 5 uA but occasionally up to 20 uA. The difficulty is that the device operates at 1.55 volts so any significant voltage drop induced by a meter destroys validity of the the reading or causes the device not to work at all.
There are $1K++ devices made to do this exact thing made for my trade but I cannot afford them. There must be a way to do this economically but I have not found it. The "uCurrent" device would help but is not available.
Is there any hope for another way?
Thanks in advance.

[fmaimon]

Build yourself an uCurrent. http://www.eevblog.com/projects/ucurrent/

[IanB]

It's pretty hard to find such a low current range on a regular multimeter unfortunately.

My inexpensive meters have a uA range but it reads up to 200/400/600 uA. Any reading below 5 uA is going to very inaccurate.

I don't know if the voltage drop should be a problem though? My meters have a current sense resistance of 100 ohms, so when measuring 20 uA the voltage drop would be 2 mV or 0.002 V. This should barely be noticeable with a 1.55 V battery--1.548 V rather than 1.550 V.

[handymandave]

Frankly, making the circuitboard and doing the surfacemount devices would involve developing skillsets I don't have time to learn.

The article that accompanies the uCurrent blog seems to suggest that at these voltage and current loads, the measurement can profoundly affect the reading.

[IanB]

The article that accompanies the uCurrent blog seems to suggest that at these voltage and current loads, the measurement can profoundly affect the reading.

No, for the majority of measurements the burden voltage or meter resistance won't be a problem. The main problem is finding a meter at all that can accurately measure down in the low uA range. Most only measure in the 100's of uA.

I guess it depends on whether you need a go/no go measurement, or an accurate current measurement? A watch repairer who once crossed my path had a clever device that measured the magnetic pulses from a quartz analog watch. He used this to tell if a watch was working when it did not have a second hand. He put the watch against the device and a light flashed if the watch was running.

[amspire]

Measuring 20uA with a multimeter is pretty easy. Just get a 10K resistor and use it as a current shunt with the multimeter on Volts.

1 volt on the DMM  = 100uA. If you use a 1% resistor, you will get a 1% accurate result.

With a 1K resistor, 1 Volt on the DMM = 1mA.

If you want more accuracy, find out the multimeters input resistance and work out exactly what sense resistor you need.

If it is 10M, then you will need (1/10000 - 1/10M)^-1 = 10,001 ohm. If you can select a 10,001 ohm resistor from a bunch of 10K 1% resistors, you will be spot on.

Richard

[EEVblog]

A good 3.5digit electronics multimeter will have a 200uA range, giving you 0.1uA resolution.
A 4.5 digit meter will give you 0.01uA resolution.
The Fluke 87V for example will have 0.2%+4 count accuracy on 3.5 digit mode, with +40 count on 4.5 digit mode.
And a burden voltage of 100 ?V/?A, so 2mV drop @ 20uA, that's only 0.13% drop @ 1.55V - good enough for your purpose.
The burden voltage mostly become an issue the closer to full scale you go, but you are operating close to zero, so not nearly as much of an issue.
The U1272A is almost identical in performance.

Dave.

[handymandave]

The article that accompanies the uCurrent blog seems to suggest that at these voltage and current loads, the measurement can profoundly affect the reading.

No, for the majority of measurements the burden voltage or meter resistance won't be a problem. The main problem is finding a meter at all that can accurately measure down in the low uA range. Most only measure in the 100's of uA.

I guess it depends on whether you need a go/no go measurement, or an accurate current measurement? A watch repairer who once crossed my path had a clever device that measured the magnetic pulses from a quartz analog watch. He used this to tell if a watch was working when it did not have a second hand. He put the watch against the device and a light flashed if the watch was running.

I have one of the blinker/beeper devices which is a big help. However, to adequately test the condition of the geartrain I need to know how much current the step motor is drawing. I can make it run fast to help loosen it up but some really need cleaning to help keep the battery life up to par. So a reasonably accurate reading is necessary. Cleaning and re-lubricating is a great deal more labor than just putting in a battery and I need to be able to evaluate that adequately while the customer is still in the shop.


Thanks AMSPIRE! I didn't know how to get there and I appreciate your advice. Now, to search for a resistor to work with my 20,000 ohm/volt multimeter.

[robrenz]

Measuring 20uA with a multimeter is pretty easy. Just get a 10K resistor and use it as a current shunt with the multimeter on Volts.

1 volt on the DMM  = 100uA. If you use a 1% resistor, you will get a 1% accurate result.

With a 1K resistor, 1 Volt on the DMM = 1mA.

If you want more accuracy, find out the multimeters input resistance and work out exactly what sense resistor you need.

If it is 10M, then you will need (1/10000 - 1/10M)^-1 = 10,001 ohm. If you can select a 10,001 ohm resistor from a bunch of 10K 1% resistors, you will be spot on.

Richard

Wont that have a relatively high burden of .2V  being 13% of the 1.55V?

[amspire]

Wont that have a relatively high burden of .2V  being 13% of the 1.55V?

Yes but a cheap multimeter can probably measure around 0.2V at full resolution. So it is the most accurate, but may be too high a burden voltage.

If 0.2V is just to much, what resistor gives a low enough burden voltage? Use a 1K resistor for a 0.02V burden voltage and most multimeters can still achieve a 0.5% resolution at 20mV. If you are measuring supply current, I would think a 0.1uA resolution would be fine. A 100 ohm resistor will give a 2mV burden voltage and most multimeters can resolve down to 1uA. The beauty is you pick whatever burden voltage you want.

[IanB]

Thanks AMSPIRE! I didn't know how to get there and I appreciate your advice. Now, to search for a resistor to work with my 20,000 ohm/volt multimeter.

Hold on a minute, you might be heading up the wrong path here. Amspire is presuming you have a modern digital multimeter. What meter do you have exactly? If you have an analog meter with graduated scale the last thing you want to do is divert current around the meter with a shunt resistor the way Amspire said.

I have an analog meter, only mine is 50,000 ohm/volt. It has a 0-25 uA range with an internal resistance of 5,000 ohms. If it were measuring 25 uA at full scale then the voltage drop would be 5000 x 0.000025 = 0.125 V. That is starting to get significant compared to 1.5 V, but maybe not too bad if we compare it with a battery at the end of its life. On the other hand with a current of 5 uA the voltage drop would only be 0.025 V, which is insignificant.

Anyway, if you do have an analog meter, you would want just to use it on the microamp range directly and just read the dial. That is as sensitive as you can get without buying other instruments.

[IanB]

Continuing the same train of thought, you might just want to purchase something like this:



Everyone these days is into modern high tech solutions, when the old fashioned, traditional way would be to use a galvanometer. Simple, cheap, reliable.

[Rufus]

but is not available.
Is there any hope for another way?

Presumably you don't need to measure a particular existing current, you could create a power supply for what you are testing and make the power supply current easily measured. I presume having a variable voltage power supply would be useful as well?

The closest standard circuit I can think of is a photodiode amplifier which turns the current flowing in a photodiode into an easily measured voltage. Your load replaces the photodiode and you need to arrange for a 1.55v (variable?) bias across the load. Not much more than an opamp, resistor and a pot. Probably run OK from a +5v supply.

[IanB]

Presumably you don't need to measure a particular existing current, you could create a power supply for what you are testing and make the power supply current easily measured. I presume having a variable voltage power supply would be useful as well?

The closest standard circuit I can think of is a photodiode amplifier which turns the current flowing in a photodiode into an easily measured voltage. Your load replaces the photodiode and you need to arrange for a 1.55v (variable?) bias across the load. Not much more than an opamp, resistor and a pot. Probably run OK from a +5v supply.

This is not an electronics design or build question (the OP is not the least bit into constructing electronics stuff), it is rather a quest for an affordable, off-the-shelf solution. I'm afraid that talk of photodiodes, amplifiers, op amps and the like is completely missing the mark.

[EEVblog]

This is not an electronics design or build question (the OP is not the least bit into constructing electronics stuff), it is rather a quest for an affordable, off-the-shelf solution. I'm afraid that talk of photodiodes, amplifiers, op amps and the like is completely missing the mark.

Especially when most cheap 3.5 digit DMM's will do the job with almost negligible burden voltage.
Burden voltage can be a real problem, but not at the low scale currents the OP wants.

Dave.