You are really trying to push your case.
When you insert a different shunt resistor into the circuit you are inadvertently altering the behaviour of the circuit.
How do you know that the circuit resistance is constant and therefore obeys Ohm's Law?
Just for example, suppose you want to measure the current flowing through a diode?

No I'm not, I already said " Of course doing that (altering the resistance dynamically) is interrupting the circuit behavior and that may not be desirable, but in cases where this is not a concern it could give us very accurate current readings, eliminating the effect of the shunt resistor entirely."

 

Are you being serious here?

Completely. On one scale you see 2.0mA and on a more sensitive scale (meaning a higher resistance, to generate a larger ∆V), you see 1.7mA. Slightly lower current in a circuit with higher resistance. That doesn't surprise me at all. And that small a difference is too little to notice visually except if they were side by side.

 

The meter shunt is usually low enough resistance that it would cause a very small inaccuracy. Typically on the 10A scale it is around 0.05 ohms. It is higher on the lower current ranges but usually, lower current circuits have higher resistance sources.
The sensible way to make an accurate measurement is to be aware of the values of the shunts in your meter and take them into account if you must have the most accurate readings. You will find that the accuracy and resolution of the readings are usually the limiting factors in most measurements.
It all boils down to knowing you instruments,.

 

Any measurement you do affects the circuit. Some old stuff, like the calibration of tube testers, a modern meter would be very wrong if it specified a meter that was 50K ohms/Volt.

Invest in an isolation transformer and Variac. Sencore made a unit with an electronic fuse, voltmeter, wattmeter and a leakage tester.

Remember, "the keep one hand in the pocket rule"

 

Thanks fellas, of course you are all correct, I'm rather rusty but as I read over the comments it becomes apparent that some basic concepts have faded from my brain!

What I suggested was interesting for me to ponder but estimating a current that would flow if the meter were not present is kind of not measuring really!

This began as I was playing around with some surplus visible and infrared LEDs that have no spec, barley any data so I wanted to see what current was flowing once it began to glow reasonably, while measuring the voltage across it and this is when I changed ranges and saw the discrepancies.

I was trying to calculate the (forward biased) diode's resistance at that point in time (it was in series with 220 Ohm resistor, about 4 volts across them).

Because I was unsure of the "actual" current I was unable to calculate its "actual" resistance - this is what got me pondering all this.

Of course I have just that one basic meter, so measuring the voltage across the diode, noting that and then using the same meter to measure the current will have slightly changed the actual voltage across the diode! This is where two meters would be helpful.

Would a "good" meter show the same kind of disparity? I guess the ability to measure very very low voltages is key here, then they can use a very low value shunt...

 

Would a "good" meter show the same kind of disparity? I guess the ability to measure very very low voltages is key here, then they can use a very low value shunt...

In my opinion, you'd be better off using your own external current shunt and use your meter strictly for voltage measurements. The meter leads in voltage mode add at least 1MΩ in parallel to your circuit (10MΩ for an expensive meter), and this is trivial to the circuit.

For example use a 1-10Ω resistor in series with your LED. At 10mA, this would give you a ∆V of 1-100mV across the shunt and something like 3V across the LED.

 

Don't sweat the small stuff. Don't over think the problem.
There is nothing wrong with the meter. What @wayneh said is a good idea. Don't switch your meter between volts and amps. If you keep on doing that you are bound to blow the fuse one day. Measure voltage across a series resistor.
10% accuracy is good enough for what you are doing.

 

What about the voltage across the 220 ohm resistor? I=V/220 Amps.

Depending on the voltmeter, the diode continuity test MIGHT show the forward voltage. Reminder, the nominal changes with the color of the LED.

In order not to disturb the circuit, you need a feedback ammeter which generally has reduced range. There is an EEVBlog device called the uCurrent that could be useful. The voltage drop using a feedback ammeter could be around 1mV. The Keithley 480 and 485 are examples. SMU' or source Measure Units can get the current up to about 1 Amp max. These devices can act as current sources, voltage sources, volt meters and current meters. 4 cables can run you $600 USD.

 

Does your meter not measure resistance? How absolutely accurate do you need the results to be? The relative accuracy between the voltages across the diode and resistor are plenty close enough to get a realistic plot of the diode characteristics. If you design a circuit using that IR diode, you will be using standard values of resistors in it anyway, with values close enough to make the circuit work, not absolute calculated values.
Experience helps you to know what you can get away with reliably. The only way to get that is by learning from the results of your experiments..

 

OK My two Fluke 8050A units arrived, only about 50 bucks a piece so definitely worth the risk.

One of the units seems very close indeed to my other devices, a small handheld DVM and my sig/gen and scope.

But the other is certainly off quite a bit so needs calibrating, anyone here done this with an 8050A? I know the procedure is documented but wanted to ask in case anyone had any tips.

What I will do is use the "working" one as my ref, I'll apply the test inputs to both units at the same time and adjust the "broken" one to match it, seems like a good first step given that the "working" unit does seem very close indeed to the other equipment.

 

I think I had one of these at work and a similar model with the lead resistance adjust. I usually grabbed the latter meter for low resistances. The only servicing I did was to replace the LCD display when they were available as spare parts. I Check out; https://www.eevblog.com/forum/repai...to-led-conversion-drop-in-replacement-screen/

The only comments is not auto-ranging and the display is hard to read at an angle. I found the model with the ohms adjust really useful.

My work portable DVM (Tek 916) had so many buttons on it, no one would borrow it. All it takes is an RPN calculator and no one would borrow that either.

 

Just for grins and giggles. Set the RELATIVE to zero it and see if it reads correctly vs another meter. Not sure how large a span it will zero... Out of the box, my used meter is good to .001V once I supplied some leads for it. Not even enough for me to even consider calibrating it to 0.

 

Just for grins and giggles. Set the RELATIVE to zero it and see if it reads correctly vs another meter. Not sure how large a span it will zero... Out of the box, my used meter is good to .001V once I supplied some leads for it. Not even enough for me to even consider calibrating it to 0.

I'll have to read the manual (I know - RTFM) because I have no idea how to set RELATIVE to zero...

 

Make sure the "RELATIVE" button is off.

Without reading the manual, it probably works like this:
Measure a voltage: Say it's 1V; Push relative; meter would read zero.
Measure 2V, it would read 1V

It would be great to set a reference for the db mode.

 

Relative is quite handy. Don't use it much on the Fluke but my 5 1/2 digit chinesium bench meter needs it quite often. Its mA scale quite often starts at 0.0017 or so. Not that I really care much about 1.7 microamps of error. Also, capacitance which the Fluke can't do.

 

The SDM3045 and the more expensive SDM3055 are in the right price range and clearly blow that utility meter out of the water.

But something else came up as I was fiddling around, I wanted to be able to measure the current through one part of a circuit and the voltage in some other part of the circuit, so I wondered if in this day and age there are multi channel bench meters, like we have multi channel scopes and so on.

To my surprise I'm not seeing any, it seems they don't exist (except as very expensive serious professional units - like 7 grand).

Is this really true? if so I wonder why? surely its easy to make a multi channel bench meter, we have 4 channel scopes after all.

It seems the only way people do this is to have two meters - very costly.

Thought?

Current and voltage measurements in different parts of a circuit are indeed tricky especially if they don't share the same circuit ground. All SDM models can do dual measurements so they are effectively 2 ch meters. There is quite a selection of the dual measurements available in the table on P41 of the manual:
https://int.siglent.com/upload_file/user/SDM3055/SDM3055_UserManual_UM06035-E02A.pdf

However for the SDM3055 the is a SC (scanner card) option that permits up to 16 individual measurements of your choice.
It's a slip in module that needs be installed and calibrated before it leaves the factory therefore unfortunately if cannot be added at a later date.
You can see a bit about it here down the page a bit.
https://int.siglent.com/products-overview/sdm3055/
Also available for the SDM3065X.

It works by using relays to connect the individual measurements to the meters circuity by in turn cycling through the # of channels you have activated and selected a measurement type for so it's not real time moreso measurement logging.