Can anyone explain these wrong uA readings on a UT61E?

[mcinque]

I've done some tests with my new DMM Check Plus.

I discovered that my UT61E is in spec, but reads wrong uA values, both in DC and AC. My UT61D is in spec in all scales. The answer is simply "Is an UT61E!"?

[sync]

The UT-61E has a high burden voltage. The DMM Check just can't provide enough voltage. Both are well known problems.

From http://www.voltagestandard.com/DMMCheck_Plus.html

1.000mA DC CURRENT SOURCE:

A 0.1% accurate active 1mA DC current source is provided to check the DC current measurement accuracy of your DMM.  The compliance voltage is approximately 500mV which accommodates load resistances from 0 Ohms to 500Ohms. The only DMM that we are aware of that is NOT compatible with DMMCheck Plus is the Fluke 116 which has a burden resistance of 3KOhms when measuring DC current.  To use this feature, set your DMM to measure DC current and connect the test leads to the provided current source terminals.  The DC current sourced by DMMCheck is guaranteed to be within 0.999mA and 1.001mA, with the actual value, as measured by our 3458A, indicated on the included as-measured record.

[mariush]

From : http://lygte-info.dk/review/Review%20UNI-T%20UT61E%20UK.html

The meter can measure from 0.01 uA to 2.2 mA in two ranges, this range and has a burden voltage of up to 0.85 volt. The actual resistance is a bit funny, most of the time it is 1 kohm, but when the burden voltage reach about 0.5 volt the resistance will drop.
Note: The resistance does not change between 220uA range and 2200uA range.

[mcinque]

Now it's crystal clear. Thank you.

[ocw]

While the UT61E is generally a good meter, Uni-Trend made two mistakes for uA measurement:
   While making uA current measurements they have a series resistance of 999 ohms--that's too high!
   They have two diodes in a reverse direction connected across that series resistor-- that causes the series resistance to change based on the current/meter's voltage drop.

Measuring current accuracy using a current source is nice, and the UT61E will give accurate readings of a current source.  However, most of the time I'm measuring the current with a short term fixed voltage connected to an unknown value fixed resistance.  Adding a 999 ohm resistor in series when the meter is being used will make the measured value inaccurate.  Worse, when those shunt diodes start to conduct across that series resistor, the value of that resistance will go down.  I measured that UT61E resistance to vary from 424 to 1,000 ohms depending on the current and therefore voltage drop across the meter.  So, you can't accurately account for the series resistance that you are adding because it will change based on how well those diodes conduct!

Due to the two problems in the meter's design I measured it to display a uA reading 5 - 10% low compared to the actual current without the meter being inserted.  That compares to the measured current source accuracy for 166 - 2000 uA typically being 0.1% or under.

mA readings just add 9 ohms resistance and only 0.01 ohms for A input readings.  While there are diodes involved with those resistors as well, I haven't seen them conduct or significantly change those series resistances.

ALL meter reviews should include measuring the meter's series resistance while in the current mode.  Good accuracy while measuring a current source is meaningless for me if it has a large series resistance while it's being inserted!

[mos6502]

Shouldn't be too much of a problem though, because thanks to the 22,000 counts, you can still measure with 1uA resolution in the 22.000mA range, right?

Or just use a 10 Ohm resistor in series with your circuit and measure the voltage drop. In the 220.00mV range, that would give you a resolution of .1uA.

[ocw]

The 220.00 mA obviously provides you with 10 uA resolution.  For lower currents your accuracy starts dropping with 1 uA resolution.

However, your accuracy is compromised more at the higher currents due to the 10 ohm resistance when reading ma current.  For example, with a fixed voltage and load I just measured 212.71 ma current with a more accurate meter.  When the UT61E was then inserted in line it read 172.46 ma (-18.92%).  The more accurate meter then read about the same thing as well.  With a linear load you can recalculate from the lower reading to account for the meter's voltage drop.  However, if the load is not linear or it requires a regulated voltage inserting a UT61E to read current with that much resistance in the meter can't be compensated for.

[mos6502]

The 220.00 mA obviously provides you with 10 uA resolution.  For lower currents your accuracy starts dropping with 1 uA resolution.

I wrote 22.000 mA ... as in, 22 milliamps. Or is the spec sheet I read wrong and the UT61E doesn't have that range?

However, your accuracy is compromised more at the higher currents due to the 10 ohm resistance when reading ma current.  For example, with a fixed voltage and load I just measured 212.71 ma current with a more accurate meter.  When the UT61E was then inserted in line it read 172.46 ma (-18.92%).  The more accurate meter then read about the same thing as well.  With a linear load you can recalculate from the lower reading to account for the meter's voltage drop.  However, if the load is not linear or it requires a regulated voltage inserting a UT61E to read current with that much resistance in the meter can't be compensated for.

Sure, you would use the 10 Ohms resistor only if you needed the 0.1uA resolution. For higher currents, you could use e.g. a 1 Ohm resistor. The point is, by using an external shunt you're not limited to the choice of 3 resistances your meter gives you, you can choose a value that best fits your voltage source and load.

[mos6502]

By the way, Uni-T seem to have learned their lesson, the UT139C only has 100 ohms of resistance on the uA range and about 2 ohms on the mA range.

[ocw]

"Sure, you would use the 10 Ohms resistor only if you needed the 0.1uA resolution. For higher currents, you could use e.g. a 1 Ohm resistor. The point is, by using an external shunt you're not limited to the choice of 3 resistances your meter gives you, you can choose a value that best fits your voltage source and load."

I want a meter which does what it claims to do, not one where you have to improvise to obtain an accurate reading that the meter is supposed to do by itself.  As previously reported, I effectively measured the current from a battery to a fixed load.  The UT61E reading was 18.92% low compared to what it was without the meter inserted.  Is that adequate?  Meter evaluations need tests like that not just the meter's accuracy reading a constant current where the added series resistance when adding the meter is irrelevant.

The added series resistance when inserting the meter went up in the UT61E as compared to the UT61D.  And adding the diodes across the meter's shunt resistors which further compromise the meter's accuracy to protect the resistors from overload doesn't seem wise.  There are better ways to do that.

[mos6502]

Well, such are the shortcomings of the UT61E. I have expounded how to get around them.

As I said, the UT139C is heaps better regarding burden voltage. It is also much better constructed in every way (PCB layout, input protection, jack design, case construction, 600V ceramic fuses, separate rubber holster). Hopefully we will see a successor to the UT61E with all these features.

[ocw]

How is the UT39C better?  It does not have a uA range.  It's mA range has a 100 ohm shunt resistor compared to 10 ohm resistance for the UT61E.  So, it's reading around 200 ma will be significantly worse than the UT61E's -18.92% accuracy.  Plus, it looks like its resistor protection diodes will start to conduct with a current like that further hurting its accuracy.  And for a fixed current source its accuracy will still be much less than the UT61E. 

The UT39E improves the regular accuracy (compared to the UT39C) but still doesn't have a uA range and still has the 100 ohm shunt resistor for the mA range.  Don't compare the UT61E's 220 uA range with the UT39's 2 ma range.  But, you get what you pay for and these are all economy meters.  The UT61E (and more so the UT39C) is not as good as those who don't account for the series resistance involved with current measurements pretend.

[ocw]

Sorry!   I addressed the UT39C instead of the UT139C.

[ocw]

It looks like the UT139C solves the UT61E's current problem by removing one digit of current resolution (and for resistance readings) as the price to pay for reducing its shunt resistance for current measurements.

[Kleinstein]

Using parallel diodes to protect the shunts from to much power it the normal way. Just if the drop can be so high, there should be something like 3 diode in series for each direction. For small currents bootstrapping of the diodes would help - but thats something for a > 5 digit meter or for very low currents protected by a large fuse.

The best way to measure small currents (especially < 1mA and autoranging) is to use a transimpedance amp. This allows a larger current range without switching the measuring resistor and even switching the resistor might be acceptable as its not that visible outside. Also the drop can be very low and diode protection is far less critical.

Autoranging in current ranges is often troublesome.