Multimeter not reading AC microvolts

[junggwok]

Multimeter not reading AC microvolts, any ideas? Will not read anything below 1 AC millivolt. AC reads fine above 1 millivolt, anything below 1 millivolt, the meter reads 0.0000

[nerdyHippy]

Have you checked the fuse?

[junggwok]

Have you checked the fuse?

Yes, fuse is fine. It doesn't measure anything less a millivolt!

[PedroDaGr8]

Multimeter not reading microvolts, any ideas?

My psychic guess is the transmorgifier is blown.

In all seriousnesd a make and model would help. Huge differences between bench and handheld, let alone between models.

[junggwok]

Multimeter not reading microvolts, any ideas?

My psychic guess is the transmorgifier is blown.

In all seriousnesd a make and model would help. Huge differences between bench and handheld, let alone between models.

Thanks for that, it is the Agilent U3402A 5.5 Bench meter.

[PedroDaGr8]

Multimeter not reading microvolts, any ideas?

My psychic guess is the transmorgifier is blown.

In all seriousnesd a make and model would help. Huge differences between bench and handheld, let alone between models.

Thanks for that, it is the Agilent U3402A 5.5 Bench meter.

Ok so it's a bench meter. These are harder to diagnose, not one I know much about.

First thing I'd check is any input protection on the voltage line. Note fuses was a bad suggestion, those are current protection not voltage. So look for things like MOVs, PTCs, etc. Check to make sure they read correctly.

[Dr. Frank]

Multimeter not reading AC microvolts, any ideas? Will not read anything below 1 AC millivolt. AC reads fine above 1 millivolt, anything below 1 millivolt, the meter reads 0.0000

That behaviour is right.
Virtually all True RMS AC voltmeters do not measure correctly below 5%  of full scale range.
Root cause is the limited capability of the RMS converter IC.
http://cp.literature.agilent.com/litweb/pdf/5988-6916EN.pdf

In 500mVAC range, the instrument may resolve 10µV, but you have to measure as close to full scale as possible.
Below about 25mV, the reading will be not precise anymore, see datasheet. Even lower inputs will give exactly zero reading, that's normal.

Joseph Geller has a very good visual explanation for the HP34401A here: http://www.gellerlabs.com/34401A%20AC%20zero.htm

Frank

[KJDS]

Here's a look at some more bench multimeters. Note that for AC current, the little Isotech was far better than the HP or Keithley.

[junggwok]

Multimeter not reading AC microvolts, any ideas? Will not read anything below 1 AC millivolt. AC reads fine above 1 millivolt, anything below 1 millivolt, the meter reads 0.0000

That behaviour is right.
Virtually all True RMS AC voltmeters do not measure correctly below 5%  of full scale range.
Root cause is the limited capability of the RMS converter IC.
http://cp.literature.agilent.com/litweb/pdf/5988-6916EN.pdf

In 500mVAC range, the instrument may resolve 10µV, but you have to measure as close to full scale as possible.
Below about 25mV, the reading will be not precise anymore, see datasheet. Even lower inputs will give exactly zero reading, that's normal.

Joseph Geller has a very good visual explanation for the HP34401A here: http://www.gellerlabs.com/34401A%20AC%20zero.htm

Frank

Thanks Dr. Frank for the clarification, it was an interesting lesson for me.

[robrenz]

This thread shows Gellers link that talks about the 34401 possibly having a firmware generated zero below a certain level. It also contains some crude testing I did on my 8846A that shows relatively linear performance down to 20µV (meaning probably no firmware generated zero.)

[junggwok]

This thread shows Gellers link that talks about the 34401 possibly having a firmware generated zero below a certain level. It also contains some crude testing I did on my 8846A that shows relatively linear performance down to 20µV (meaning probably no firmware generated zero.)

Interestingly, my Keithley 2000 showed a similar performance to your Fluke meter, seems to have linear region right down to about 30µV.

[Dr. Frank]

This issue did not let me alone.

I have found out, that even the big brother, the HP3458A, uses the same AD637 for its analog ACV mode.
In this mode, it has the same resolution (6 1/2 digits) and similar accuracy figures, like the 34401A.

I have measured various precise AC voltages @ 1kHz from a Fluke 5200A AC standard by the display readings from the HP3458A and the HP34401A, both in the 1V ACV range.
And I also have measured the converted, analog DC output at pin 6 of the AD637 in the HP34401A, just before the signal is fed into the A/D converter. So there is no further possibility to analogously manipulate the signal.

The diagram clearly indicates, that the AC converter in both instruments delivers linear readings down to about 0.5 mV, but then saturating at a residual value. Therefore, smaller values are unprecise and not useable any more.

The 34401A display reading deviates from the linear transfer function below about 4mV (0.4% of full scale).
 
That clearly proves, that in this and similar HP / agilent instruments, (e.g. the U3402A), the software artifically sets readings to zero.

Such AC converters are able to deliver results about 10 times smaller than that, but it is questionable, if that region is really useful, due to decreasing accuracy. 
It's better, to have lower ranges available, like the HP3458A, which provides an additional 10mVAC range.

Frank

[robrenz]

Thanks for checking that out Dr. Frank.    Your chart basically agrees with the one from Gellers info.  I wonder what the 8846A is using for the RMS converter?  If the same converter then maybe they are characterizing the non linear response area in software closer to zero.

[Andreas]

Hello,

I did a similar measurement as Frank with different Instruments.
(in 100mV Range with 100 Hz sine wave against Frank 1V and 1kHz).

Just to compare some instruments for AC-Capability.
- HP34401A (100mV range)
- K2000       (100mV range)
- PS5444A   USB-scope, calculated RMS-value, range as required. 300kHz bandwith limited (similar to DMM bandwidth).

Input was from signal generator of PS5444A with 2*12 Ohms as load (gives around 10:1 divider with the internal 50 Ohms).

The 34401A has its deviation beginning at 0.5% of range (similar to Frank)
whereas the Keithley and the high resolution scope have very similar values down to 0.1% of range.

With best regards

Andreas

[Kleinstein]

The classical RMS chips like AD637 are limited in the low range. So these meters may not work well at very low voltages. Some meters have a more or less hidden auto-ranging to compensate for this, so that lower values can be read as well.

The HP3458 / 3457 and some other newer meter with TrueVolt RMS should not use an analog RMS chips, but digital sampling. They may work reasonably well down to lower values, but there can be still larger errors at very low readings.