Are all Autorange Handheld DMM's Usless for Current .. Almost?

[amspire]

I have been living in the dark ages a bit.  On my bench, I have 3 Fluke 8050A non-autoranging multimeters, and the main handheld DMM I have been using has also been non-autoranging. I actually like selecting the range I want, and the readings are instant.

I have several autoranging DMM's but it struck me I never use them for measuiring current.

In my recent post on the Amprobe AM-270 multimeter, I was a bit shocked to see a 1.65V burden voltage at the top of the 500mA range. The reason was of course is that they use one resistor for the 500uA and 5000uA ranges, and one for the 50mA and 500mA ranges.

So if a meter needs at least 100mV to get full range at a 50mA current , then it will have a 1V burden voltage at 500mA.

Are all the autoranging DMM's bad like this? Are there any good ones? It seems to me that they should just have the switch select all current ranges explicitly, and forget about the fiction of autoranging for current.

If I buy a Fluke 87 and measure 400mA on the 400mA range, I will have a burden voltage of 720mV.

If I buy a $10 multimeter with a manual ranging switch, I will probably be able to measure the same current with a burden voltage of less then 200mV.

Go figure!

Richard.

[IanB]

You've inspired me to measure some of my meters. I'll report back later when I have the results.

[amspire]

Here is the reference multimeter against which I will judge all other multimeters:

http://www.aliexpress.com/product-fm/494791754-AC-DC-Professional-Electric-Digital-Multimeter-Tester-Checker-Free-Shipping-Wholesale-wholesalers.html



$3.09 including shipping.

About 200mV burden voltage on the 200mA range.

One buyer feedback of 5 stars out of 5. 

The big guys should be able to do better then that, surely?

Richard.

Edit: That $3.09 price was for a minimum order of 5 multimeters for $19.75. Goes all the way up to $3.95 including shipping for just 1 multimeter.

[amspire]

Just did a bit more testing and it looks like even the non-autoranging meters are not great.

I was wrong in assuming the cheap meters would have a 200mV burden - it looks like they are more like 0.8V burden.

My Fluke 8050A's are about twice as good as the autoranging meters, and by using the 2A range that I usually use for power supply current testing, I get a 0.2v drop at 400mA with a 0.1mA resolution.

It seems to me that all that is needed is a preamp for the current ranges, and they could keep the burden voltages below 0.1V easily. The lower heat in the sensing resistor would probably mean improved accuracy figures.

So are there any brands that take current measurement seriously?

Richard

[PeteH]

Have you seen?:

http://www.eevblog.com/projects/ucurrent/

[IanB]

All of my meters have a resistance of about 2-3 ohms on the DC mA ranges and 100 ohms on the DC uA ranges.

On the cheap meter you pictured, the resistance was 2.7 ohms on the 200 mA range. That would give a burden voltage of about 500 mV at 200 mA.

The RS 22-812 has a burden voltage of 800 mV at 400 mA, and the Protek 6500 1.5 V at 500 mA. Others are in the same general range, at 2-3 mV/mA, with the RS 22-812 being the best out of those I measured.

The Micronta 22-204A is 1000 mV at 500 mA FSD, so apparently DMMs are neither better nor worse than analog meters.

[amspire]

Have you seen?:

http://www.eevblog.com/projects/ucurrent/

Lucky for Dave, Fluke is too dumb to add this to their $400 multimeters.

Considering the amount of effort companies like Fluke have pored into multimeters, I am struggling to see how they can justify burden voltages of 0.7 and up. Dave has proved that solving the problem is not impossible at all.

[IanB]

When I measured the current draw of the RS 22-812 meter (950 uA) I got exactly the same reading when I used the uCurrent and when I used the DMM directly on the uA range. With a meter resistance of 100 ohms the burden voltage in that case would have been 95 mV. Considering the battery was about 9.5 V the burden voltage was insignificant.

[Zad]

I don't know of any handheld meters that do it "properly" (via a transimpedance amp) or even anything approaching the uCurrent's performance. I was surprised to see that my new bench meter, a Tektronix DMM4020 (a rebadged Fluke 8808a) does the low current modes (200uA, 2mA)  with a proper transimpedance conversion. Then again for £500 I guess it should do. Even then, it has the following specs on the higher ranges:

20mA burden <0.05V
200mA burden <0.5V

2A burden <0.1V
10A burden <0.5V

These values aren't massively better than a good hand-held. I guess the high current range has a 0.05 Ohm resistor, and the low current range uses 2.5 Ohms.

[PeteH]

I think the fuse adds a lot in terms of burden.

Dropping this protection feature in the uCurrent leaves only the terminals, switch, routing and sense resistor. I had a quick look at a 5x200 (500mA) fuse and the burden at 0.5A (rated) is 285mV.

Quite a heavy burden; but we expect no less in a robust multimeter...

[amspire]

Peter you are probably correct.

If you need to have a sub 1A fuse, then it has to have a signifigant resistance just so it can work.

Now if they used the 20A sense resistor for mA, and added a preamp like Dave uses, then the protection fuse is a 20A fuse and the fuse resistance is way lower.

So I guess the question this:

Is it better to have a 500mA fuse (or whatever the meter uses) on the mA ranges and put up with a big burden voltage, rather then a 20A fuse.

Is this an issue of protecting the equipment under test rather then the multimeter. So if you put the multimeter mA current range across a power supply in a device under test, it can probably easily blow a 500mA fuse, but it may easily not blow a 20A fuse.  The results could be blown devices or a damaged mains transformer in the device under test.

I know the guys at Fluke are totally brilliant, and I am just trying to grasp the logic on why the current ranges have to be so bad. None of this explains why the high uA range has to be bad - that is not a fuse issue. Just bad design.

Richard.

[PeteH]

The sense resistance is proportional to the sensed current range, so a trade-off must be made in terms of the fusing. If your fuse is too large for a given input range, the sense resistor will take the majority of the voltage drop (and hence a large amount of power dissipation).

The majority of testing I have done with current meters has relied on the power supply providing the limiting element... I think it is not the responsibility of the multimeter to protect your DUT, as you could easily select the 10A range and as a result end up giving the multimeter permission to no longer protect your mA range product.

I would agree with you Richard, Fluke would have incorporated low burden sensing if it was economical (but then again, we're talking about Fluke...) or if it did not sacrifice reliability.

[IanB]

I imagine the 500 mA fuse is there to protect the meter. If you put 10 A through (say) a 0.1 ohm sense resistor the power dissipation would be 10 W. There would be smoke.

I was amused and saddened at the number of complaints about the RS 22-812 meter from reviewers that said "the fuse is too sensitive--it keeps blowing when I use the meter". If there was no fuse these same reviewers would be complaining that the meter was too fragile and broke as soon as they used it.

The answer is not clear. A sensitive device in capable hands will easily be broken in clumsy hands.

[alm]

The fact that low burden voltage appears to occur mainly in bench meters, which have usually less protection features than hand helds, suggests that protection may be the hard issue. A higher resolution and accuracy helps, since it allows you to use a higher range. Nothing prevents you from just setting an autoranging meter to its highest range, which results in a lower burden voltage.

[Zad]

A low burden meter does not necessarily mean reduced protection. If you do it properly with a transimpedance amplifier, the amplifier inputs should be protected with a pair of forward biased low leakage diodes. Under fault conditions, the op-amp can deliver perhaps 20mA, at which point the input burden rises to 0.6V and the diodes conduct. Your trade-off is then diode leakage current versus maximum forward current, this will dictate the type of fuse used.

If you are trying to use a micro-current range to read 10A then you must expect the fuse or other protection system to operate!

[IanB]

A low burden meter does not necessarily mean reduced protection. If you do it properly with a transimpedance amplifier, the amplifier inputs should be protected with a pair of forward biased low leakage diodes. Under fault conditions, the op-amp can deliver perhaps 20mA, at which point the input burden rises to 0.6V and the diodes conduct. Your trade-off is then diode leakage current versus maximum forward current, this will dictate the type of fuse used.

If you are trying to use a micro-current range to read 10A then you must expect the fuse or other protection system to operate!

Perhaps you could explain a little more, as I am not familiar with the details you describe?

Above it was mentioned that the resistance of a 500 mA fuse introduces a high burden voltage, so it was suggested to use a 10 A fuse on the low mA ranges.

The situation now arises that 10 A can flow continuously through the meter without blowing that fuse (unless some electronic protection intervenes first).

What form would the current path take through the meter between the input terminals, the 10 A fuse and the sensing circuitry?

If this current path could withstand 10 A continuously without overheating, what would it consist of?

If this current path included electronic over current protection, what would protect the protection circuitry from high voltages across the input terminals? (Without re-introducing a high burden voltage...)

[ejeffrey]

A transimpedance amp is probably only good up to 2 milliamps or so -- the low power opamps you would want to use in a handheld device can't usually source much current and if they could it would drain your battery pretty quickly.  There is no reason you couldn't make a meter with transimpedance for the microamps range and a ucurrent style amplifier for milliamps.  In a bench meter you could use a high current opamp or a discrete transistor output driver and run in transimpedance mode up to a few hundred milliamp, although you would have to watch the TCR of the gain setting resistor, as it will self-heat quite a bit.  Obviously you would still use a shunt + amplifier for the amps ranges, but that is much less of a problem.