How do you calibrate your gear.

[chinoy]

So I had a Fluke 105B where the buttons stopped working. And my kitchen foil hack also stopped working.
So I got a Hantek 5000 series Scope. The measurements did not match my old fluke scope. So I got a new Fluke 223A.
Now every measurement I take is different. The new fluke does not match the old fluke. And the scope matches nothing.

I know there are labs where you can send in your gear to calibrated.
Also know that the new Fluke 223A can be re-calibrated.

My interest lies in measuring very low voltages only i.e. all under 1 Volt.
You know the kind of  voltages you get off a tape head.

Are there any tricks or easy ways to confirm my meters accuracy ?.
Like maybe buying a new battery or building some kind of circuit that puts out a very accurate voltage I can use to calibrate my gear.

[tggzzz]

More information needed....

Numbers, not adjectives.

Define measurement techniques.

Compare results with specifications.

Pictures showing "incorrect" results.

[Psi]

You can't compare a scope voltage measurement to a DMM measurement.

A DMM is designed to accurately measure voltage.
A scope is not, in fact many scopes only have 8bits of vertical resolution, although 10bit is getting common.

So if your scopes vertical scale is set to say 1V/div and your scope's screen has 8 divisions, and the trace is in the middle, then you can measure +/- 4V.
The 8bit vertical resolution means your +/- 4V is represented by a number between 0 and 255, and that is not much at all

[chinoy]

The new Fluke DMM says 6000 count. Could you explain that in simple english.
I really dont understand why pictures. Im measuring a point on the PCB. I need to turn a trim pot till I get 0.45 volts
If every meter is giving me a different reading Im not sure which one to trust.
A friend told me just go with the new Fluke as they would have done some kind of calibration before shipping it and its the newest of all the gear you have.
I have some other 6000 count no name meters but even here every reading is different.
On the scope I press the measure button and it throws up RMS voltage which is what I'm taking as the value. And confirming it with the scale Im running at i.e. 100 Mv per division. Even if I could just figure out of the meters is the most accurate Id be happy.
There must be some circuit or trick to get an accurate number. I even tried 3.3 Voltage regulators but even voltage regulators have a margin of error. i.e. the out put changes from IC to IC.

[alm]

First, realize that to calibrate means to get a piece of paper (or PDF file) that says whether the unit was within set specifications. Be sure to actually specify if you want adjustments.

Every decent test instrument will come with specifications like the displayed value should be within 0.9%+0.5%, which means 0.9% of the reading (e.g 1.5V) and 0.5% of range (e.g. 6V). Or 0.9%+5 digits, which is 0.9% of the reading + 5 least significant digits.

If you measure the same signal with two meters in parallel, and both meters are within their specifications, then the intervals of the minimum and maximum actual values calculated from their displayed reading plus and minus the uncertainty calculated from the accuracy specifications should overlap. The actual value should be in the overlapping range.

Be sure to pay attention to little footnotes like "only valid after zeroing", "only valid from 5%-100% of full scale", "only valid for sinusoidal signals from 40 Hz to 500 Hz" etc.

If after this exercise the meters are not overlapping, then one or more will need adjustment. If they are overlapping, then you might be expecting too much. A brand new Fluke meter should be well within its specifications, so I'd trust that (within its specified accuracy).

Check the manual for adjustment procedure. While there are some affordable standards like the DMM check plus that can be used as a sanity check, adjusting a DMM generally requires a range of signals from 100 mV to 1000 V DC and AC, etc. Some instruments allow you to just adjust a single range, other require going through all ranges and functions before they will save the adjustments.

[xrunner]

The new Fluke DMM says 6000 count. Could you explain that in simple english.

The counts is the maximum number that can be displayed by the meter display, in your case it's "6000". But what does that mean? It depends on the range the meter is using. It could mean 6000 volts, or 600.0 volts, or 60.00 volts, or 6.000 volts and so on.

A 3000 count meter would do the same ranging - could mean 3000 v, or 300.0 v, or 30.00 v etc. There are many explanation on the all-knowing internet. 

[tggzzz]

The new Fluke DMM says 6000 count. Could you explain that in simple english.
I really dont understand why pictures. Im measuring a point on the PCB. I need to turn a trim pot till I get 0.45 volts
If every meter is giving me a different reading Im not sure which one to trust.
A friend told me just go with the new Fluke as they would have done some kind of calibration before shipping it and its the newest of all the gear you have.
I have some other 6000 count no name meters but even here every reading is different.
On the scope I press the measure button and it throws up RMS voltage which is what I'm taking as the value. And confirming it with the scale Im running at i.e. 100 Mv per division. Even if I could just figure out of the meters is the most accurate Id be happy.
There must be some circuit or trick to get an accurate number. I even tried 3.3 Voltage regulators but even voltage regulators have a margin of error. i.e. the out put changes from IC to IC.

Numbers, not adjectives, please.

Of course they all read different values. That is to be expected.

You need to understand the specifications of the various instruments, in particular their tolerances. For each instrument, take the reading and work out the range of possible values. Compare those range of values. For example, instrument #1 might read 1.24V, and instrument #2 1.27V. If the tolerance specification for #1 is 1.2V to 1.3V, and for #2 is 1.25V to 1.3V, then both those readings would be consistent.

Why pictures? So we can see important things you didn't tell us, e.g. measuring AC voltages, not DC, and measurement techniques.

[TimFox]

The new Fluke DMM says 6000 count. Could you explain that in simple english.

The counts is the maximum number that can be displayed by the meter display, in your case it's "6000". But what does that mean? It depends on the range the meter is using. It could mean 6000 volts, or 600.0 volts, or 60.00 volts, or 6.000 volts and so on.

A 3000 count meter would do the same ranging - could mean 3000 v, or 300.0 v, or 30.00 v etc. There are many explanation on the all-knowing internet. 

6000, 3000, etc. counts is a more specific statement than the previous "3-1/2 digits" display.  Usually, "6000" means "5999", but that is understood.

[ebastler]

Im measuring a point on the PCB. I need to turn a trim pot till I get 0.45 volts
If every meter is giving me a different reading Im not sure which one to trust.

When others here have been insisting on "numbers please", they mean: What specific values are your various meters displaying when you measure the same voltage? And have you compared the actual difference to the tolerances stated in the meters' datasheets?

You cannot expect the meters to agree to the last digit on the display. Your meters' datasheets will state tolerances which amount to at least several counts on the last digit, and quite a bit more in some ranges -- take a look. So even if you were to tweak a potentiometer inside each meter to make them read exactly the same today, for a specific voltage, at a specific room temperature -- expect them to differ tonight when it gets cooler, or a few days from now when the potentiometer has settled in.

Another angle: How precisely do you need to adjust that voltage, e.g. the 0.45V you mentioned? There must be a tolerance on that adjustment target too.

You can buy voltage references to check your meters at home, and for some models of multimeters there are documented ways to adjust them internally. But unless you have very specialised needs for high-precision measurements, or are a "Volt nut" who considers high-precision metrology a hobby in its own right, you probably don't need those calibration tools.

[bdunham7]

My interest lies in measuring very low voltages only i.e. all under 1 Volt.
You know the kind of  voltages you get off a tape head.

Start by stating some reasonable (numerical) expression of what you expect that voltage to be.  AC, DC or both?  Presuming you are looking for an AC voltage, what voltage, frequency and source impedance do you imagine you have?  Is it a sinusoid, noise or something else?  If not a single frequency, bandwidth?

Now look at the specs of whatever you are trying to use to measure that voltage.  Does it have sufficient BW?  What are the specified accuracy tolerances? 

How much difference are you seeing between different measuring instruments?  There are many reasons why you might see differences and we'd need to know all the details to even have a hope of answering your questions.

[pickle9000]

Get a process calibrator.

[807]

...I need to turn a trim pot till I get 0.45 volts
If every meter is giving me a different reading Im not sure which one to trust...
 
...I have some other 6000 count no name meters but even here every reading is different...

As other members have mentioned, it's normal for different meters to give different readings. The point is, how different are these readings? Give us some actual measurements. Just saying that they are different has no significance. If you set the pot to give 0.45v on one meter, what readings do you get on the others?

[rdl]

You might not want to spend the money but there are cheap voltage references available on eBay and probably on AliExpress.

[bdunham7]

To sort of answer the OP's question with an example, I set up the following small bench experiment:

I set up an AWG (SDG2042X) to put out 450mV_RMS @ 5kHz plus the 15th harmonic @ 900mV_p-p and combined that with 50mV of noise with a 20MHz BW.  Oddly the channel combine feature didn't work as I expected, but that's a different issue--it displayed on a scope as you'd expect but an an amplitude different than I get adding them all up. 

I then measured that signal with every DMM within reach and with the scope (SDS2354X-E) as well, using the StdDev function.  Here they are rounded to 100µV for consistency and suppression of irrelevant digits.  No surprises IMO.  On meters with separate mVAC and VAC ranges, I tested both as they did vary on some models.

Siglent SDS2354X-E:  310.5mV

Fluke 8506A, Hi-Accuracy mode: 310.6mV

HP 34401A: 307.6mV

Fluke 8808A: 307.5mV

Fluke 8842A: 307.1mV

Fluke 289:  307.0 (both mVAC and VAC)

Fluke 189:  297.2mV (mVAC), 0.3042V (VAC)

Fluke 116:  ~300mV (mVAC, very unstable), 0.225V (VAC)

Fairchild 7000A (AVG responding): 0.2832V

Fluke 27 (AVG responding): 280.0mV (mVAC), 0.282V (VAC)

and (I almost forgot it) my old friend the Fluke 98 Scopemeter, after some careful setup....310.4mV (multimeter mode, 100µs/div, 300mV range) w/ ~ 1.0mV peak noise.  It also displayed a reasonably stable 0.310V in the 1V range at the same timebase setting.

[CJay]

Man who has one watch knows the time, man who has two watches is never sure.

Seriously though, the big question is does it need to be absolute accuracy or just repeatably inaccurate?

I.E. if it indicates 0.45V when it's actually 0.47V or 0.43V does it make a damn of difference as long as the error is always the same?

[chinoy]

You might not want to spend the money but there are cheap voltage references available on eBay and probably on AliExpress.

Thank you that is the best lead I got so far.
I will sit down and list the values I am getting From each instrument. The next time I build one of these circuits. For now I have just used the newest Fluke I have.

[JohanH]

When measuring small signals, particularly if they are high impedance circuits, the instrument you use will load the circuit, which affects readings. How much depends on the input impedance of the instrument, but also of your circuit. An oscilloscope has 1 MΩ to 10 MΩ of input impedance. That's like loading your circuit with a resistor of the same value. A modern cheap voltmeter typically also have around 1 MΩ of input impedance, whereas more professional ones have impedance measured in gigaohm (GΩ). Check the specifications!

As an example, I've built my own voltage references and I get different readings with the 34401A meter depending on if measured with its default 10 MΩ impedance mode and when changing to 1 GΩ (it sucks that the gigaohm mode can't be turned permanently on with this device). It's AC input is 1 MΩ in parallel with 100 pF.

[EPAIII]

If you are going to compare the readings on two meters, they should both be connected to the test Voltage at the same time.

To sort of answer the OP's question with an example, I set up the following small bench experiment:

I set up an AWG (SDG2042X) to put out 450mV_RMS @ 5kHz plus the 15th harmonic @ 900mV_p-p and combined that with 50mV of noise with a 20MHz BW.  Oddly the channel combine feature didn't work as I expected, but that's a different issue--it displayed on a scope as you'd expect but an an amplitude different than I get adding them all up. 

I then measured that signal with every DMM within reach and with the scope (SDS2354X-E) as well, using the StdDev function.  Here they are rounded to 100µV for consistency and suppression of irrelevant digits.  No surprises IMO.  On meters with separate mVAC and VAC ranges, I tested both as they did vary on some models.

Siglent SDS2354X-E:  310.5mV

Fluke 8506A, Hi-Accuracy mode: 310.6mV

HP 34401A: 307.6mV

Fluke 8808A: 307.5mV

Fluke 8842A: 307.1mV

Fluke 289:  307.0 (both mVAC and VAC)

Fluke 189:  297.2mV (mVAC), 0.3042V (VAC)

Fluke 116:  ~300mV (mVAC, very unstable), 0.225V (VAC)

Fairchild 7000A (AVG responding): 0.2832V

Fluke 27 (AVG responding): 280.0mV (mVAC), 0.282V (VAC)

and (I almost forgot it) my old friend the Fluke 98 Scopemeter, after some careful setup....310.4mV (multimeter mode, 100µs/div, 300mV range) w/ ~ 1.0mV peak noise.  It also displayed a reasonably stable 0.310V in the 1V range at the same timebase setting.

[Swake]

If you are going to compare the readings on two meters, they should both be connected to the test Voltage at the same time.

Absolutely not. JohanH just explained in the message previous to yours one of the principal reasons why not.

[PlainName]

If you are going to compare the readings on two meters, they should both be connected to the test Voltage at the same time.

Absolutely not. JohanH just explained in the message previous to yours one of the principal reasons why not.

Johan explains why the test voltage may differ depending on instrument. EPAIII explains how to circumvent that - the absolute value isn't relevant, only the difference between the two displays.