This very probably has already been suggested somewhere upthread, but measure some other low value precision resistor, such as 10 Ohms 1%, to see if it measures in spec. If it does, those 3.9 Ohm resistors are badly out of spec., or more likely, you were ‘mistakenly’ shipped 10% tolerance parts.
For small resistance values (<10ohms), I prefer to use higher current (100mA, 200mA,1A etc) from a DC power supply or even a battery.
Measure the current and the voltage directly in the resistor leads, then calculate the R=V/I.
This way you eliminate any cable or contact resistance from the measurement.
Generic multimeters uses very small current levels which generate very small voltages. That's why the measurements are not accurate for lower resistance values and are accurate for high resistance values.
In addition, multimeter includes the multimeter cables and contacts in your measurements.
Think if you need to measure 0.1ohm resistor. No generic multimeter can do it.
But if you inject 1A, you will get 100mV across the resistor.
1A and 100mV are easily and accuratelly measured by any multimeter.
Measure the current and the voltage directly in the resistor leads, then calculate the R=V/I.
This way you eliminate any cable or contact resistance from the measurement.
Generic multimeters uses very small current levels which generate very small voltages. That's why the measurements are not accurate for lower resistance values and are accurate for high resistance values.
In addition, multimeter includes the multimeter cables and contacts in your measurements.
Think if you need to measure 0.1ohm resistor. No generic multimeter can do it.
But if you inject 1A, you will get 100mV across the resistor.
1A and 100mV are easily and accuratelly measured by any multimeter.
Hi ron68,
The accuracy of his meter isn't good enough, plus to avoid errors, you need two meters. One for current and the other for voltage.
Way further ahead to simply get a good meter with much better high frequency response. The next step would be to measure the voltage across the driver or load resistor, and if your meter has poor frequency response and accuracy (as this one does), it will lie to you. These meters were designed to do electrical work, not electronics.
The accuracy of his meter isn't good enough, plus to avoid errors, you need two meters. One for current and the other for voltage.
Way further ahead to simply get a good meter with much better high frequency response. The next step would be to measure the voltage across the driver or load resistor, and if your meter has poor frequency response and accuracy (as this one does), it will lie to you. These meters were designed to do electrical work, not electronics.
The point of my test is not confirm or establish the resistors' value. If the measured value does not change (roughly in half) as expected, then move for a different way to measure your resistor value.
adason,
I have a dats and do not find it it to be consistent. It is less than one year old.
I do all of the calibration stuff and it still gives strange readings with low values of resistance.
Its ability to measure inductance is very unreliable - the readings are usually ridiculous.
I bought it to measure low value resistors and feel I wasted my money. I do not doubt it does fine for measuring driver parameters.
I have a dats and do not find it it to be consistent. It is less than one year old.
I do all of the calibration stuff and it still gives strange readings with low values of resistance.
Its ability to measure inductance is very unreliable - the readings are usually ridiculous.
I bought it to measure low value resistors and feel I wasted my money. I do not doubt it does fine for measuring driver parameters.
Just do it the proper way. I worked in calibration labs for electronic test instruments, so I am trying to offer solid practical advise.
Everyone wants to do things well, cheaply. I am a fan of saving money, but there are times (most of the time actually) when it pays to get the right stuff and do things the right way. Industry does things certain ways for a couple important reasons. It's cheaper, and you get the right answer. There is nothing more expensive than having the wrong answer. You don't need the documentation part of why industry does certain things, but you still need the correct answers.
Everyone wants to do things well, cheaply. I am a fan of saving money, but there are times (most of the time actually) when it pays to get the right stuff and do things the right way. Industry does things certain ways for a couple important reasons. It's cheaper, and you get the right answer. There is nothing more expensive than having the wrong answer. You don't need the documentation part of why industry does certain things, but you still need the correct answers.
What has frequency response to do with measuring the DC resistance of a resistor? If you want to measure the impedance of a part, you need a LRC bridge or so, a very different kind of equipment.
Hi Havoc,
Absolutely zero. Nothing at all. But consider his other applications.
Anyone buying a meter to just measure resistors as a hobbyist is really silly. You will use the meter for everything. To measure frequency response, you also typically measure AC voltage. What is the accepted audio range please? Remind me. Now look at the specs of his existing meter. Hmm, doesn't cut it, does it?
I took his other uses into consideration when I made my suggestions. Since I have practical work experience in calibration of test equipment, and much in audio where I use the same instruments, and in telecom where I maintained paging systems (audio again), I'm trying to take everything into account when I made a recommendation.
Note I suggested buying specific used meters known for accuracy and the ability to hold their calibration (many meters don't). I could have recommended a new Keithly, Keysight or Fluke. I have some new ones in use, so practical experience again. The Keysight 34465A has some neat features good for audio, the 34461A less, but still excellent. For basic stuff, the 34401A is about the best bench meter ever made. An industry standard and there are many out there.
I'm thinking long term with meters that will cover the most uses. Therefore, instruments that offer the highest value per dollar spent.
Absolutely zero. Nothing at all. But consider his other applications.
Anyone buying a meter to just measure resistors as a hobbyist is really silly. You will use the meter for everything. To measure frequency response, you also typically measure AC voltage. What is the accepted audio range please? Remind me. Now look at the specs of his existing meter. Hmm, doesn't cut it, does it?
I took his other uses into consideration when I made my suggestions. Since I have practical work experience in calibration of test equipment, and much in audio where I use the same instruments, and in telecom where I maintained paging systems (audio again), I'm trying to take everything into account when I made a recommendation.
Note I suggested buying specific used meters known for accuracy and the ability to hold their calibration (many meters don't). I could have recommended a new Keithly, Keysight or Fluke. I have some new ones in use, so practical experience again. The Keysight 34465A has some neat features good for audio, the 34461A less, but still excellent. For basic stuff, the 34401A is about the best bench meter ever made. An industry standard and there are many out there.
I'm thinking long term with meters that will cover the most uses. Therefore, instruments that offer the highest value per dollar spent.
Hi,
I think it's interesting to always have 2 cheap multimeters, so the person can use the procedure I mentioned (measure DC current and voltage).
My suggestion is simply use a better scale (more current through resistor) and better method, which eliminates the need of considering the leads contact or cable resistance.
Instead of a super expensive equipment capable of measuring submilivolts or any other complex method, the idea is to generate higher voltages to read (in tenhs or hundred of milivolts), which can be easily done with reasonable accuracy by a regular cheap multimeter.
Hi ron68,
I agree with you, but the OP didn't say he had more than one meter. That's why I commented you would need two meters to minimize errors.
If you have a great deal of experience, you can figure out ways to minimize errors and figure out your uncertainties. That is what a metrologist does (not a weatherman). You figure out the currents and voltages to enhance the accuracy of your instrument, and also your test setup. You would use a Kelvin connection for your method.
As I said above, if the OP has to purchase something else, may as well make it serve as many purposes well as possible. Additionally, a skilled person can use less good equipment to get good answers sometimes. They won't because it is a waste of time.
Never fight your equipment. It just isn't worth it.
I agree with you, but the OP didn't say he had more than one meter. That's why I commented you would need two meters to minimize errors.
If you have a great deal of experience, you can figure out ways to minimize errors and figure out your uncertainties. That is what a metrologist does (not a weatherman). You figure out the currents and voltages to enhance the accuracy of your instrument, and also your test setup. You would use a Kelvin connection for your method.
As I said above, if the OP has to purchase something else, may as well make it serve as many purposes well as possible. Additionally, a skilled person can use less good equipment to get good answers sometimes. They won't because it is a waste of time.
Never fight your equipment. It just isn't worth it.
Yes, what I sugested is in fact the 4-wire Kelvin connection, where we separate the current flow from the voltmeter wiring.
And this is interesting cause it is generic for any resistor range.
Imagine if you need to measure the resistance of a car alternator or starter coils, where the resistance can have extreme low values such as 0.01ohm.
With a two US$50.00 regular multimeter, you can measure it with good accuracy (0.5%).
Injecting 5A, you just need to read 50mV, which is easily done on the 0-200mV scale, for example.
And this is interesting cause it is generic for any resistor range.
Imagine if you need to measure the resistance of a car alternator or starter coils, where the resistance can have extreme low values such as 0.01ohm.
With a two US$50.00 regular multimeter, you can measure it with good accuracy (0.5%).
Injecting 5A, you just need to read 50mV, which is easily done on the 0-200mV scale, for example.
Hi Ron,
That's the concept. Low resistance measurements are not easy to get right with high accuracy. Even 1/2%.
But you have to take into account inaccuracies of both meters at the range and scale they are on. This is different from the basic DC accuracy. It also depends on how long since the last calibration and if they were even in tolerance out of the box (many are not). For the current monitoring meter, heating of the shunt element may also add to errors as it warms up (depending on current). Ambient temperature may also shift readings. Self heating of the resistor you are measuring may also add to uncertainties.
For most people, just buy a solid, good industrial brand of resistor and accept it is tolerance. It may be close to the edge of tolerance, but your measuring equipment and method may show it well out, or well in tolerance. Such are uncertainties in measurement. In short, measuring it didn't tell you anything unless it is way out and your measuring setup and equipment are shown to be much closer to reality. If you are intent on measuring these things, either take a metrology course and set things up properly, or just buy equipment that will do the job with low uncertainty.
I'll say one thing, it isn't a bad idea to study metrology if you're into electronics anyway. You will learn a lot that applies to many situations beyond measurements. You'll also know when your instrument is lying to you.
That's the concept. Low resistance measurements are not easy to get right with high accuracy. Even 1/2%.
But you have to take into account inaccuracies of both meters at the range and scale they are on. This is different from the basic DC accuracy. It also depends on how long since the last calibration and if they were even in tolerance out of the box (many are not). For the current monitoring meter, heating of the shunt element may also add to errors as it warms up (depending on current). Ambient temperature may also shift readings. Self heating of the resistor you are measuring may also add to uncertainties.
For most people, just buy a solid, good industrial brand of resistor and accept it is tolerance. It may be close to the edge of tolerance, but your measuring equipment and method may show it well out, or well in tolerance. Such are uncertainties in measurement. In short, measuring it didn't tell you anything unless it is way out and your measuring setup and equipment are shown to be much closer to reality. If you are intent on measuring these things, either take a metrology course and set things up properly, or just buy equipment that will do the job with low uncertainty.
I'll say one thing, it isn't a bad idea to study metrology if you're into electronics anyway. You will learn a lot that applies to many situations beyond measurements. You'll also know when your instrument is lying to you.
There is nothing wrong with that voltmeter if you know and understand the limitations. It's ok to check DC voltages, AC voltages on mains and low frequency sine waves, etc. I have several about like that one. Nothing wrong with it. But in this case it isn't the right tool. However not because of the bandwidth. That is why people longer into electronics have several instruments according to what has to be "measured".
AC voltages really need a scope because then the waveform becomes just as important as some equivalent RMS value.
And don't let yourself be fooled by digital meters with very high specs. Remember audio is about sound. And the speed of sound, the most important variable in this whole game, has variations like 10%...
Hi Havoc,
My experience with most meters not made by Fluke, HP, Agilent, Keysight is that they do not hold their calibration. Most are not in tolerance new and many can never satisfy accuracy specs as advertised. So the first assumption made is that the meter in question is in fact "in tolerance". Not a reasonable assumption unless you have had a level 6 calibration done (records all values, not just out of tolerance values). Most meters use individual resistors in the divider and gain scale sections. They do not track thermally and often drift in all directions (having had to correct calibration before with older meters). The better meters use a ceramic substrate and laser trimmed resistors for those functions. This also lowers stray capacitance and helps keep frequency response flatter without weird peaks and dips. High frequency compensation adjustments are made "closed case" hopefully. Stray capacitance from the case and pressure really throws things off (try calibrating a Fluke 87 for example).
Yes, you should monitor with a scope, many are 2% accurate for Y values and suffer the same issues as meters with individual resistors setting scaling and dividers. However if you are close to a sine, an RMS meter works great. High frequency response will composite waveforms with harmonics (non-sine) and yield accurate results. Your scope will show peak values, not RMS and that's only with a good probe that is equalized properly. A meter with poor HF response will not catch the harmonics properly and yield errors even it the fundamental is 1 KHz, common in audio measurements. With power line stuff, we are interested in 50/60Hz and their harmonics, so the Amprobe would be okay there. Not for audio, not even close.
Knowing the limitations, as I mentioned in another way, is key. But most people don't know and have to study the subject to understand, as I've said.
Music or control signals, it's all the same. Nothing special about music and we were talking direct current and sines. Where does music come in? Musical waveforms adhere to all laws of physics, and are composed of a collection of equivalent sine waves. No electronic component knows whether a signal is noise, music or whatever. It matters not. So if you want to go down that road, we're into audio analysers now. Those are extremely expensive. Still they show electrical qualities of a waveform, whether it is music or noise doesn't matter.
My experience with most meters not made by Fluke, HP, Agilent, Keysight is that they do not hold their calibration. Most are not in tolerance new and many can never satisfy accuracy specs as advertised. So the first assumption made is that the meter in question is in fact "in tolerance". Not a reasonable assumption unless you have had a level 6 calibration done (records all values, not just out of tolerance values). Most meters use individual resistors in the divider and gain scale sections. They do not track thermally and often drift in all directions (having had to correct calibration before with older meters). The better meters use a ceramic substrate and laser trimmed resistors for those functions. This also lowers stray capacitance and helps keep frequency response flatter without weird peaks and dips. High frequency compensation adjustments are made "closed case" hopefully. Stray capacitance from the case and pressure really throws things off (try calibrating a Fluke 87 for example).
Yes, you should monitor with a scope, many are 2% accurate for Y values and suffer the same issues as meters with individual resistors setting scaling and dividers. However if you are close to a sine, an RMS meter works great. High frequency response will composite waveforms with harmonics (non-sine) and yield accurate results. Your scope will show peak values, not RMS and that's only with a good probe that is equalized properly. A meter with poor HF response will not catch the harmonics properly and yield errors even it the fundamental is 1 KHz, common in audio measurements. With power line stuff, we are interested in 50/60Hz and their harmonics, so the Amprobe would be okay there. Not for audio, not even close.
Knowing the limitations, as I mentioned in another way, is key. But most people don't know and have to study the subject to understand, as I've said.
Music or control signals, it's all the same. Nothing special about music and we were talking direct current and sines. Where does music come in? Musical waveforms adhere to all laws of physics, and are composed of a collection of equivalent sine waves. No electronic component knows whether a signal is noise, music or whatever. It matters not. So if you want to go down that road, we're into audio analysers now. Those are extremely expensive. Still they show electrical qualities of a waveform, whether it is music or noise doesn't matter.
Oooohhhhh... Now you are going into very deep waters here 😉
Rest assured, we are on the same wavelength here. But what you wrote in that last paragraph is against all audiophile lore. Beware you do not end up tarred and feathered.
I welcome the attempt to tar and feather!
To somehow believe that audio signals are exempt from the laws of physics pretty much explains everything for those folks. lol! Components do not have memories or precognition.
Audio lore is big business. Truth is what those people fear.
To somehow believe that audio signals are exempt from the laws of physics pretty much explains everything for those folks. lol! Components do not have memories or precognition.
Audio lore is big business. Truth is what those people fear.
Picking up on the suggestions of using a DC supply and measuring the voltage. Would the diode setting on the meter give similar results? It won't display resistance but it should show if they are a match.
The diode function should work, assuming the test current is accurate. With HP/Agilent/Keysight it is 1 mA. You still run into lead contact and resistance issues. That function is only intended to show a working diode junction and not a qualitative answer. Plus most meters have limited voltage compliance on diode check.
In other words, it is a functional test only. Don't trust it.
In other words, it is a functional test only. Don't trust it.