Those DMMs give me the creeps.

[thaumaturge]

Crow isn't my favorite food, but here goes...
I stand humbly corrected. Going back over a group of schematics I find "most" "high end" meters do indeed use feed through DC blocking caps on their AC circuits. It is still not, however, a feature that I would expect to find on a cheap handheld meter.
Doc

[Mooly]

Originally Posted by tvrgeek

All above true, but you need a load across the transformer.

Why ?

[mrmodemhead]

Originally Posted by thaumaturge

Crow isn't my favorite food, but here goes...
I stand humbly corrected. Going back over a group of schematics I find "most" "high end" meters do indeed use feed through DC blocking caps on their AC circuits. It is still not, however, a feature that I would expect to find on a cheap handheld meter.
Doc

It's been my experience that if a "low-cost" multimeter has only two AC volts ranges, 200V and 700V, it's a dead giveaway as to what is inside. It will be a cheap & cheerful ICL7106-based design, AC input will be DC-coupled (no blocking cap), it will use a simple half-wave rectifier with a single diode, and it will exhibit the exact behavior that pellesand described previously. For some reason the digits "830" often appear in the model number, regardless of the branding. This design must be printed on the back of every restaurant menu in Shen Zhen.

But once you get into the next price bracket, say around $30-50 US, the design will probably include a DC blocking cap and a precision rectifier, so AC performance improves quite a bit. Of course, I wouldn't just assume that without checking though.

[tvrgeek]

Originally Posted by Mooly

Why ?

Otherwise the meter is the load. Without a external load, you have no way to get the effect of internal winding resistance or other parastatic elements. Meters have a relative high impedance but they differ, so a load of less than the meter will then provide reliable results. 100K would do in this case. You don't care what the winding produces into air, you care what it will do in a circuit.

[Mooly]

Originally Posted by tvrgeek

Otherwise the meter is the load. Without a external load, you have no way to get the effect of internal winding resistance or other parastatic elements. Meters have a relative high impedance but they differ, so a load of less than the meter will then provide reliable results. 100K would do in this case. You don't care what the winding produces into air, you care what it will do in a circuit.

All the meters in my tests were run together in parallel.

The AVO 8 mk 5 has a sensitivity of only 2000Ω/V on 10vac ranges and above, 1000Ω/V on the 10 Vac range and on the 3 volt ac range it draws 10ma for F.S.D. so in the tests that include the AVO the load is actually very very "high".

Running the meters in parallel ensures all are reading the same source voltage irrespective of loading.

[Atilla]

hm, sorry that I haven't reported back, but work has been crazy lately.

I haven't dug into them yet, but I performed what seemed to me to be the simplest test I could do. I took a 390ohm 0.1% precision resistor I had and measured it.

DMM 1 says 388, so I'm going with 'totally unreliable', if it can't even measure resistance
DMM 2 says 390 so at least that part is fine.

DMM3 is at work, unfortunatelly, so I need to bring them together and see why does it show so much higher AC voltage, compared to DMM2. It's a small, auto-ranging device, which should be reasonably good in theory, but I'm starting to suspect that DMM2, the most generic device is the one that's properly done.

Which of course, doesn't mean that it's AC circuit is better than what mrmodemhead describes

[Mooly]

When I was checking the AVO over, I measured and recorded the readings that my Escort DVM produced for the precision resistors. The tolerance of these resistors is impressive, from memory I think they were marked as 0.1% even on the high value ones.

AVO precision resistors as measured on DVM.

2K16 reads 2.161

177K reads 177.0

800K reads 8.03

1M8 reads 1.796

4M [resistor 1] reads 4.01

4M [resistor 2] reads 4.02

8M reads 8.05

[tvrgeek]

Pretty good I'd say Mooly. Most handlheld meters are about 1% +/- a digit or two. If you want better than that, better get a five terminal bridge in a environmentally controlled oven. That'll run you about $100,000.

Atilla, Meters actually only measure volts. Two volts to be precise. All other scales are based on additional components. For resistance, that means an internal current source and reference resistor. Resistance is not the simple measurement you may think. Battery and calibration, temp, switch contact repeatably and so on. Metrology is actually a very difficult science. We should be amazed we do as well as we do. When I started as a lab tech, we would think ourselves very lucky to adjust anything to 5%. Half our 260's did not even have the mirror scale to help with parallax, and when handheld, the mass of the movement caused errors depending on meter position. Life is good.