Reference voltage for calibrating voltmeters

Maybe your AC line? There are plenty of precision voltage references out there for pennies, but that's not "everyday" I suppose.

Compare it to another multimeter. If you don't have one, go to a store (Radioshack, etc) and ask to try one of theirs out. Just take a small battery and compare the two readings.

How do you know the other meter is not faulty? Even new things can be at fault.

Make one from a fixed voltage regulator say 5 volts and 12 volts. If you want a very accurate voltage source for a reasonable price, get a good transformer and bridge rectifier, then add some suitable smoothing capacitors say 10,000uF times 2. next use 24volt first stage then 18 volt then 12 volt then 9 volt then five volt. the 5 volt will be rock steady and you have several other useful voltages as well. If you can make this you will always find it able to help you out. There are plenty of useful data sheets on the net. So you have no excuse for not having an accurate voltage source.

A 2000 ohm pot in the center leg of a 3 terminal regulator will allow you to increase the voltage from the named voltage and calibrate it.

They are normally good to within 1/4 volt. If you select from a bunch you will find one close to the named voltage.

So a +5 can be trimmed to a 6 or 9 or 12 volts this way

Oh dear me The whole idea of a reference is that you can't change the fundermental, it must be fixed and absolute or you can not calibrate against it can you?

Think about your idea what is wrong with it?

You have a measuring instrument of unkown accuracy now you want to find out how good it is so you present it with an as yet unkown voltage. God words fail me.

common 3 term regulators are only good to within .25 volts. This is hardly a standard!!

Selected ones might be right on. A small variable resistor in the ground leg can be used to trim those below voltage upwards to the precise voltage, or a higher voltage. They call this making a secondary or even tertiary standard, depending on how directly it's being checked against a primary standard.

Thus a standard can be made that you can take home to check your meter with. The other sources are the electrochemical standards they sell, which are portable primary standards. (primary in that they generate an EMF on their own from chemical soluutions and are used within temperature and current drain limits)

Ok but you will never know unless you have an absolute supply to begin with. Any uncertainty calibration adding padding will just add to an already confused state.

Yes I know not every regulator is absolutely within 0% off its published value.

BUT how will you know whether to increase or decrease the voltage unless your meter is calibated to begin with? Oh you of little brain power.

Try a half cell for reference. That is what we used to use in calibrating meters in the meter shop. We used to check and verify the calibration of AVO analog meters as part of the annual re-certification required in any precision lab.

you can create your own half cell by using two known metals such as copper wire and zinc immersed in an electrolyte.

The galvanic potential of metals remains nearly constant and any change is based on known deviations such as temperature, purity of the electrolyte etc.

In our shop secondary reference we used a silver - silver chloride half cell. It acts like a precision battery. However, you must be careful not to load it with a low impedance meter or you will deplete the half cell and it becomes inaccurate.

Your car battery is not sufficiently controlled to be a known accurate reference. Theoretically it should read 12.6 V

The 13V reading is due to a surface charge on the plates. In time this would dissipate. Unfortunately lead acid batteries also suffer from self discharge, so it is not usable as a reference over long time spans of several months.

Far too complex and messy.

• What kind of accuracy do you need?
• Is it AC or DC that you need to calibrate?
• If AC, what frequency?
• Max and min voltage?
• You said "multimeter" what about resistance and current, do you need to calibrate them as well?
• do you need certification?

Before you answer these questions, all the answers you got, to say the least, are partial. In order to achieve what you need, you will first have to come up with this information.

Wangito.

First of all define the accuracy you are looking at.

1%, 0.1%, 0.01% or 0.001% of the set voltage?

For 1% zener of 5.1V may be OK and for greater precision look at www.national.com www.linear.com www.analog.com www.maxim-ic.com etc to select the right reference voltage device.

The thing is the absolute accuracy is relative to the application but repeatability the ability to always indicate the same result for the same measurement is just as important.

You are explaining the difference in precision and accuracy. Measurement precision does not guarantee accuracy but provides data that can be tried for accuracy and it may be accurate or may not be accurate. You may require signal averaging filter or fitting to equations or none may work if your measurement is only white noise. In that case it is a great achievement as white noise source is not easy to make. You try to land on moon Mars but end up on Jupiter. Something like that.

two load cell sensors plus a display (mVdc) assemly with hydralic cylicders to make a strength test machine, (tensile), how this machine is calibrated? using load cell one to calibrated against another one?

thanks

RS Components (usual disclaimer) offers a calibration service for multimeters. For a fee, the meter can be calibrated against Standards and a certificate issued.

In the field of industrial instrumentation, particularly during the commissioning phase, evidence that the test equipment was in calibration during the commissioning period establishes a high level of confidence that the plant will operate as designed.

It is even more important in the case of a pharmaceutical plant, as the test certificate is a vital part of the Installation Qualification and Operational Qualification phases of plant construction.

Yeah seen the cost factor your meter will have to be worth it. Min £35 up to £120.

I believe new Lithium-ion button cells have a fairly precise and stable output voltage if not heavily loaded. You can construct a multi-point reference with a number of these cells connected in series if it's just the DC portion of your multi-meter that you want to calibrate. This should suffice unless you need laboratory measurement accuracy.

It depends on required accuracy. If 10% is enough, You can use battery. But battery voltage decreases with time (selfdischarging: within a year, NiCr batteries 3..10years), and depends on temperature - main reason of day to day instability.

For 1% accuracyZener diode is OK as reference. For 0.1% ..0.01% special reference diode is reqiuired. Or You can use Weston cell if Your multimeter has High input impedance.

If You need 0.0003% use Fluke 732A

EASY WAY: USUALLY MULTIMERERS DO NOT CHANGE THEIR INDICATION MUCH PER YEAR. SO CHECK YOUR MULTIMETER ONCE/YEAR IN NEAREST ELECTRONIC SHOP/SERVICE POINT.

IF YOU WANT TO BE SURE, BUY ANOTHER ONE AND CHECK ONE AGAINST ANOTHER WHEN YOU NEED SURE RESULTS. THIS WILL BE PROBABLY THE EASIEST AND CHEAPEST WAY.

czgut

Well, Fluke calibrator is pretty expensive.

To the list of questions: user must be sure what is his/her meter. Presently hand-held DMM (Digital MultiMeter) is so popular that its Data Sheet is often missing. You may anyway note:

(A) How many digits are on its display. E.g. four will indicate at worst 1% inaccuracy if first (significant) digit could be 0 or 1 (3 & half type) or 0.5% if it can indicate 0-1-2-3 (3 & 3/4 type). The other way is to look at max scale. For the first type they could be: 0.2 or 2 0r 20 etc Volts. 3 & three quarters will have max e.g.: 400 mV-4V - 40V

(B) Running Technical College Lab I used to calibrate /check-outs DMMs the Data Acquisition internal board (installed in the computer) and software e.g LabVIEW by National Instruments Austin TX.

Pozdrowienia dla bogatego wlasciciela Fluka. Nie mam polskich czcionek do druku na CR4, sadze, ze Pan mnie jednak zrozumie. Byly gdanszczanin.

..re Fluke .. I was joking, but is still one of best choices if You need 0.0003% accuracy.

re (A) YES

They are usually based on ICL 7106 like (TC <80ppm/deg, zero <1uV/deg, scale <5ppm/def, noise <80uV, linearity <1count, rollover<1count) giving pretty good 0.1% Vdc basic accuracy at full scale, going down to 1% at 0.1 of full scale. But usually real accuracy is lower , some 0.3% ...0.5% full scale, due to input divider errors, and other errors..

But You can buy them for some $10 which is good price for that performance.

Reliability: if such a multimeter has 3years expected lifetime, giving <0.1%/day failure rate, then having extra multimeter You will get <0.0001% probability that both will fail tomorrow. So paying 10$ for 1000fold improvement in reliability is not a big price.

Pozdrowienia dla Poludnia z Polnocy (Greetings to South(ern123) from north)

czgut

Are you two talking to each other in your own language??

What is it Polish? Slovakian?

John.

Yes.. in polish. We just started to get aquainted today through CR4. We have found to be from the same Town and University..

Wesolych Swiat Wielkanocnych = Happy Easter to You John!, And to all CR4 Readers!

Czesiek

A few statements are in Polish. Nothing coded. Just Hallelujah greetings for Easter.

By the way I have read some of your remarks, Dear ElectroMan - they are so GOOD - almost directly instructing everybody without needs to read books or to pay consultants.

I have used British, even pretty old, video tapes in Mechanical Engineering lectures and all of US: students and lecturers found them extremely simple (comparing to our college books) but not simpler! giving US necessary well designed information /knowledge transfer Channel.

In my lab I use a Schlumberger precision voltage standard.

It covers the range from 1 uV to 100 Volts by thumbwheel switches, its accuracy is specified as 20ppm.

Its in our constant temperature calibration room...

For everyday use check your multimeter on any old battery, if it indicates approximately what you know it should then the calibration should be fine - assuming its within its calibration date.

Modern multimeters don't need calibrating often unless you have dropped it or tried to measure mains voltage / car battery voltage on the current ranges etc...

As for the advice above using series regulators forget it... You either want a very good calibration source or a rough check to make sure its working i.e. a battery @ roughly 1.5volt a cell.

John.

I suggest the use of trimmed series regulators for those with no ready access to such equipment that you have. A small board and a 12V ac adaptor can be made into a reasonably accurate device. A resistor and zener can be added for other voltages. Selecting the zener is also required if you want better than off-the-shelf results

Yes, fair enough Aurizon, but the temperature coefficient is not very good, same with long term stability...

and how is he going to calibrate the calibration source?

I was just pointing out that most modern multimeters of 3½ digits or more need accurate calibration yearly, and in between are remarkably stable so only need the occasional check to make sure it works.

John.

Yes, the temperature stable internal references have pretty much killed off the half cell business. True comments, but good anough for most users.

The annual requirement for calibration of 31/2 DMM of modern making is largely an artifact of adminsitrators driven by standards associations who remember the good old days (and who want the calibration fee business). Unless the DMM has an 'event', it should stay within specs for year and most of the new ones do a self test and warn of anything wrong.

IN fact with modern self levelling/self testing gear many of the calibration rituals of yesteryear are gone with the wind or should be. They still live on in government procurement though.

Dear Arizon

Good point about zener. 5.1V zener has the lowest temperature drift and both lower and greater voltages drift more.

Sorry Shyam, but the lowest temperature coefficient zeners are right at the changeover from zener action to avalanche which is at 6.2 volts...

If you look at the standard references of decades ago (1N821 series) they were made from internally a 5.6 Volt zener / avalanche and a series diode to compensate the drift...

By carefull selection the 1N829s could be selected to drift less than 10ppm /*C

Then along came the LM3999 (I think) and you had a precision factory trimmed 10.00 volt reference....

John.

Dear John

You may be perfactly right. I looked at those figures some 35 years ago so those are from old memories. LM199 etc were made to those 6.95V zone perhaps exact cross over point may uncertain at that period as zeeners were made from different levels of dopings.

I think you have current information and good that you pointed it out. Perhaps I am more to 5.1V as it works with current days 5V circuits easily.

Probably 0V/deg TC crossover is near 5.1V. Diodes in 821..829 series have 5.6V voltage, but have no zero TC. They have some +2.3mV/deg. Added in series diode polarised in forward direction has -2,,,-2,5mV/deg TC coeffocient (dependent on current). By adjusting current it is possible to get zero ppm/deg TC at certain temperature. Temperature dependence of such a diode is V(T) = a + b*T +c*(T-T0)^2 around that temperature.

Czesiek

Actually they have a voltage of 6.2 volts because they use a 5.6 volt zener with a diode in series.

As I said in my post.

I agree with You! 5.6V on "Zener diode" + 0.6V on in series forward diode gives 6.2V.

But I understood discussion between You and Shyam as: At what Voltage of avalanche/Zener diode TC is near zero.

regards!

Czesiek

There are many businesses that must use calibrated multimeters; e.g., those doing electronics contract work. They usually have their meters calibrated (at least) annually by a local calibration shop. The calibration shops charge ~$25 to verify the calibration (if no work/repair is required). If you want to avoid that expense, I suggest you take a new 9v battery to a shop with a calibrated meter and ask them if they can read its voltage for you within 1%. They probably won't charge you for that. Then, you can check the battery with your meter against the value they read. It's only one datum point, but it will give you an idea if your meter is reading correctly.

take a new 9v battery..

..and in the meantime You will heat it in Your pocket(glider pilot method to keep battery working, when You are >2km above ground and battery and You are getting colder 0.7deg/100m), or in car by sun, and voltage will change..

..but it is a question of required accuracy..

czgut

You were right. I overestimated influence of temperature on battery voltage in my comment to Your post. I have measured it and found to be 0.06%/deg. Sorry!

czgut

You could use a 5.1 Volt Zener Diode Test Circuit......you would need al least a 6 Volt source to turn on the 5.1 Volt zener.......say 4 dry cells......and use a 1000 ohm to 470 ohm resistor in series to limit the battery current........

The 5.1 VOLT ZENER is the best choice of zeners as it has a practically zero temp coeffiecient........that is it will not drift as it warms up!

When you get the circuit hooked up.........take it to a friend to measure the actual zener voltage as manufacturers usually have a 5% tolerance on the 5.1Volts.......you will pay more for a 1% tolerance.......

Sincerely.........Jim Walters

I would add some 2..4 more batteries, to get lower dependence of Output voltage on battery discharge.

And use 1N 829 reference diode, and adjust current so, that TC =near 0ppm/deg (but for this we need to have another circuit for reference)

czgut

Your Meter could be accurate. To charge a battery, a Higher voltage is used. Many Regulators in cars typically charge with 14 V peak ripple. The accuracy of the alternators are adjustable and rarely result in exactly 12 V charge in the battery.

Thanks for all the replies.

For the sort of accuracy I need it looks like a fully charged battery will be near enough . but I do not know the exact voltage of the battery.

Nominally it is 1.5 volt dc (or 9 volts) but do rechargeable nicads and hydrides etc, have different voltage outputs than throwaway types. Otherwise it means using an 'unknown' voltage to check the meter.

It is like the old story of the time-keeper at the local factory - he used his pocket watch to time the hooter at 8.00am and at 5.00pm. When he retired he was presented with a clock (no pun meant) which was to be collected from the local clock-shop.

Chatting to the clock-shop owner, the old time-keeper mentioned that all his working life he set his watch at 7.45 am by the clock in the shop window every morning as he walked past on his way to the factory.

The shop-keeper said that was interesting, because every evening he set his clock against the 5.00pm factory hooter.

Horace, I've heard that one before, and it never ceases to make me laugh....

Being involved in measuring and calibration I sometimes see almost exactly the sort of mistakes made in businesses, some times in large well known businesses that really should know better....

John.

I think the worst one just recently was by a large household name company that had a problem so I investigated...

They had to hold a test pressure very accurately at 0.500 bar, they insisted that it was very accurate as the test guage was regularly calibrated by an approved lab...

When I checked the 'test guage' was little more than a standard dial bourdon tube guage which had a full scale range of 10 bar!!!

They had been setting the pressure to only 5% of the scale length of a 2.5% accuracy guage - which as we all know is only accurate to 2.5% for the middle 2/3 rds of its intended full scale range....... INCREDIBLE!!!!

John.

Hi Horace40!

What voltage do You want to measure (for example 1.5V, from 1V to 10V)?

What accuracy You would like to have? 1% of measured value? 0.1% of measured value?

What probability, that You are not wrong, You want to have: 99%? 99.99%? 99.9999%?

This would help us to help You

Having meters and gauges we encounter time to time needs:

(a) Check-out for the simple functionality (just does it work? - indicate?). One of the expression of this activity is called: One - Point Check-Out. For this task we have used Canada Made chip that has pretty stable and accurate 10.00 Vdc output. Input to the chip was DC Power Supply with adjustable 0 to 30 Vdc

(b) Analog type indicators are the subject of International or Regional /Industry Standards. Very popular set of checking points in so called five -point method is:

10 - 30 - 50 - 70 - 90 % of full range / scale. Process transmitters with current output 4 to 20 mA dc would be checked at 5.6 mA -8.8 mA - 12 mA - 15.2 mA - 18.4 mA.

There are in use also: 0-25-50-75-100 % 5-point method, and three point (different settings usually dictated by the industrial applications.

DMMs are usually pretty stable, resistance elements inside the meter are very accurate too. Hand-held units maybe checked for one point ones per year or as needed (suspected malfunction).

As was previously stated, we need to know the accuracy of the required calibration.For calibration standard certification, the source must be at least twice as accurate as the instrument to be calibrated.

A temperature-stable reference voltage can be made by using the differential voltage between 2 precision zener diodes of different values.Even though the temperature induces changes, the differential voltage will remain virtually unchanged, because temperature affects both equally.

Notice I said "reference voltage" not calibration source.The best way is to send it to a certified calibration company, and build and test your reference voltages immediately thereafter.That way, you can spot-check periodically for a need for recalibration.

Hope this helps.

Batteries do not have a constant voltage it depends on curent and level of cherge use better a good zeener diode.

This is a dead topic..... nobody knows what accuracy is required and what accuracy the meter is needed to be...

So it makes sense to wait for the questioner to return.... if ever!!!

John.

look at #28 - partial answer

A brand new "1.5V" battery is probably around 1.6 V.

Why not try an AC/DC adapter with a voltage switch. Granted, you don't know how accurate the settings are, but the manufacturer has probably taken some steps to make it reliable! It sounds like it might be enough for you.

I am indebted to all the helpers. But somehow, although more knowledgeable, I am not much further forward. Somewhere down the line a 'Test House' must have a reliable reference voltage sources - otherwise they will not know how accurate their 'standard' meters are. Understandably, each test house (down the line) would have more sophisticated equipment to ensure their accuracy - for instance if temperature (of a zener diode say) is a critical, then their thermometers themselves need calibration against a known temperature reference - and their ammeter will need calibrating - and so on.

So back to me, in order to avoid complex calibration (and delay and cost). A simple check like a dry battery cell would do (if I knew the true voltage output) - because with three different multimeters I get three (marginally) different readings from the same battery. Admittedly my LCD DMM is potentially more accurate (and probably even more accurate because I don't have to 'guess' analogue readings between divisions).

Basically I come back to the simple question what trust I can place in my meter readings. I have a 50ft narrowboat (see Avatar logo) with a 12 volts power system - so measuring dc volts accurately is sometimes quite important especially some distance from the batteries due to volts drop etc. It would just be useful to know I can rely on my meter. Hence my question about reference voltages.

I like the idea of the zener diode - but I need someone to draw the circuit please.

OK, a zener is like a hole in the side of a tank. You increase the water level and zero flows until you reach the hole. Then water exits the hole and it stops rising. You now have a controlled water level as long as you have some exiting the hole and if you think of the hole as infinite and it will never let the water level above the hole.

A zener is analogous to the hole. You connect the zener to the voltage source and you increase it until it starts to conduct electricity 'down the hole', as it were.

This is the zener 'knee' or the regulation voltage. a 5.1V zener has a knee at 5.1 volts and as you try to saire the voltage higher, it passes it to ground...within limits.

A voltage and a current passing through a zener = a wattage if you use volts and amps. So if you want 5.1 volts and you need to pass .1 amp then 5.1 x 0.1 = 0.51 watts. You should then pick a 1 watt zener in case you need to pass more amps. Now the voltage that flows through the zener can also flow through your circuit as a parallel resistance to the zener, but you must make sure you do not have such a high drain that the voltage falls below the 'knee', and is then not regulated.

Dear Aurizon

Looks you are good at zener theory, then can you explain, what makes zener potential to be at a particular voltage and why we can not have other potentials. You also have to explain the factor that makes sharp curve compared to some Zener Diodes that are not so sharp.

Re: battery as voltage reference: I have measured 1.5V (nominal voltage) Sony alkaline battery, partially used, in two temperatures: at 20.75'C +- 0.5deg Voltage was 1.483568V by 8 1/2 digit multimeter (10ppm accuracy), and at temperature 27.7'C +-0.5deg was 1,48264V. That means 0.000928V voltage change due to 7deg temperature change. This is equivalent to 0,0625% change of voltage and makes 0,0075%/deg temperature coefficient of battery voltage.. I think this is fair enough accuracy to your purpose. So probably You can benefit from Your idea to use battery as voltage reference. 9V or 12V would be better than 1.5V. But as the battery voltage decreases with time, it is better to have reference Voltage based on good reference diode.

Re. trust in reading: Using typical LCD DMM at home You can assume that it's Vdc reading uncertainty is below 1% with probability 97%. Probability of error >2% is below 1%.

If two of Your multimeters read the same source voltage with difference <0.5%, then You can assume they are correct to 2% uncertainty with probability >99.99%.

When used at boat - (salt water and higher humidity), use two multimeters, or multimeter and reference voltage, and assume 99% probability that readings are OK if both DMM read near the same value (difference <1%). This is rough estimation of confidence of readings.

You are right: at sea multimeter has at least 10 times higher failure rate than at home, and double checking is recommended. And it is better to use Instruments dedicated to use at boats. They are better preserved against influence of salt and humidity.

czgut

Horace40 wrote: Basically I come back to the simple question what trust I can place in my meter readings. I have a 50ft narrowboat (see Avatar logo) with a 12 volts power system so measuring dc volts accurately is sometimes quite important especially some distance from the batteries due to volts drop etc. It would just be useful to know I can rely on my meter.

REPLY - Arrgh! why didn't you say you were looking for boat voltages!!?

In such a situation it really doesn't matter much if you are off 10 millivolts or more. All the calibration discussions here has to do with when you need micro volts accuracy and long term stability. In boats you really don't care about the micro volt differences and unless you are doing corrosion prevention work, millivolts are not that important either. Forget the tables that relate battery voltage to charge levels as a percentage of full. That is a very poor and inacurate method. Various factors such as ambient temp, duration since last charging session and even how recently you applied a load to the battry will drastically affect the voltage reading.

Consider a voltage reading on the battry to be no better than 10% acurate for capacity purposes. It makes the foregoing discussion about less than one pecent accuracy rather a moot point.

you mentioned checking voltage loss. Except for during commissioning check this is a pointless exercise. I'm an electrical sysem designer for large yachts and I do get involved in measuring voltage drop in actual installations. The better way to get realistic readings is to use a very long test lead and measure the voltage difference between two points. This is a more realistic and true real time measurement of voltage drop at that exact instant in time. Since load current typically varies over time and cables heat up with current, taking measurements at first one place and then at another is futile. you have no idea how much change in temperature or current or wire resisance occurred in the interval.

If you are trying to see how much charge is remaining, a much better method is to use one of seveal battery capacity meters on the market. Since you have a narowe boat I would assum eyou are either in the UK or perhaps northern Europe like in Holland.

Xantrex has a product called Link 10. Mastervolt has a similar product that even looks the same. Last tim eI cheked they also called it a link 10 ( they copied it) and Victron has a BMV 501 which does the same thing functionally.

These meters use a micro processor and very acurate 41/2 digit meter to measure instantaneous current through a precision shunt and simultaneously the voltage input and voltage drop drop acros the shunt. The micro-processor computes the amount of amp hour energy consumed since the battery was last fully charged. It subtracts each amp hour unit from the pre-programmed maximum capacity of the battery. yuo can set alarm points to warn you when approaching the lower safe limit of discharge. Because battery capacity varies in direct relation to how great a discharge current is applied, the micro processor also applies Peukertrs formula to compensate.

Apart from that application you really do not need to know the exact voltage closer than a volt in most marine trouble shooting work. forget the annual calibration crap, never mind the 0.0001% stabilized and temp corrected accuracy. buy a cheap meter. Whe you dro pit in the bilge or overboard yo uget to buy a new one in any case. After dropping my new Fluke RMS meter that way I smartened up and seldom bring a precision meter o nboard any more. I carry two spare meters in the van for backup and the occasional time I really do need an accurate meter. Hope that clarifies some boating application volt readings. If you are in fact trying to do corrosion detection work, you really shouldn't be given your evident unfamiliarity with meters. Corrosion work does involve millivolt readings BUT!!! it also relies very much on comparative readings with reference to a silver -silver chloride half cell and comparison with the galvanic tables. Here again temperature and water salinity plays a very significant part in the actual readings. I should warn you that corrosion work is quite specialized and relies on a great deal of experience. Other boat volt measurement can be done with plus minus one volt tolerance and stil give good diagnostic reults

Very good selling (and teaching) pitch!

I am looking at a digital multi meter that would suit your needs made by a company called TENMA list number 72-7925 available from www.farnell.co.uk their part 918-3663 at £19.21 + VAT. Very good value. It has all you will ever need. Look it up. There are pleny of more expensive that have better functions but none you will require. All the best. It will be cheaper and quicker than building a voltage source.

If you can find a mercury battery, it is a very stable voltage source.The voltage remains steady till the very end of its life.They were frequently used in calibration equipment because of this characteristic, and also in cameras.They have been outlawed in the USA, but may be available elsewhere.

My question is why do you need such accurate measurement for your application?

A cheap meter from wal-mart is ok for general purposes.If you are looking for voltage drop along a length of wire, simply measure it at your source, and then at your destination (under load) .The absolute voltage at either end will not matter, only the difference from end to end. A large voltage drop (difference) indicates a loose connection, or an undersized conductor for the applied load, or leaky insulation on the wiring.

IF you want to build a very accurate, cheap reference source, simply use 2 zeners of different values, say a 12v and a 6v and wire them in parallel with the appropriate current limiting resistors, and the differential voltage between them (6volts in this hypothetical case) will not vary with temperature.Be sure to mount the zeners very close to each other so that they experience the same temperature.This should be accurate enough for your purposes.

Hope this helps.

HTRN

I appreciate the replies and although I use my meter on 12vdc circuits in my narrowboat, I posed the original question because it occurred to me that taking costly action based on the readings of a 'cheap' multimeter was not the way to go.

For example, I have an inverter to give 240v 50Hz pure sine-wave ac -and it works fine - except when running the microwave oven (pulling 100 adc) when the TV picture shows signs of interference (nothing serious - just a nuisanse for a rew minutes) I checked this with the suppliers who said it was probably volts drop and I need a bigger cable to feed the inverter - and if that did not work - then get back for more advice. Nice one!.

Well I thought, rather than rewire the boat, it would be more scientific to check the volts first - and then I got thinking - what about my meter- was that accurate - and then further thought to how to check it, and how 'reference' voltages are produced. And then to the question how to do it myself.

It looks like a zener circuit will do the job.

Thanks all for your help.

Horace40 wrote:I have an inverter to give 240v 50Hz pure sine-wave ac -and it works fine - except when running the microwave oven (pulling 100 adc) when the TV picture shows signs of interference (nothing serious - just a nuisanse for a rew minutes)

REPLY

one of the characteristics of Victron inverters is they behave like an inductive load. As battery voltage declines, they pull more current, which in turn produces greater voltage drop, .. . and so on. For that type of situation the only realistic way to measure voltage drop is to have the long lead and measure from one end of the positive lead at battery end to the other end at the inverter (load). Under load the positive cable will exhibit a voltage drop so the battery end will be slightly more positive than the inverter end. At 100A draw this can be as much as a volt. For any cable run longer than 2 meters I use locomotive cable in the 3/0 or 4/0 sizes. Otherwise the drop becomes excessive and you get problems such as you describe.

In N. America some locomotive cable suppliers now have ABYC certification. Sorry but I am not familiar with who has CE certification for installations in the UK.

Inverter generates noise. Oven also generates high frequency noise. Both of them can influence TV picture. If noice goes to TV by cables, filter on cable supplying TV can help. If by electromagnetic waves - screen. In general: good one point grounding helps keeping noise low.

I think above mentioned noise is more probable cause of Your problem, than voltage drop on cables.

Checking against voltage drop: Before rewiring boat, check if better cable will help: Make temporary connections, (outside intallation channels) with more thick cables, and check if this will help.

You can calculate voltage drop on cables with 100A current: dV=1,75V*length[m] / cross area [sq mm]. For example AWG 8 = diam 3.78mm = cross area 7.94 sq mm , length 1m will give 0.22V drop on supplying cable and 0.22V on return cable, toghether 0.44V = 3.6% of 12V = not a problem. But 10m length means 4.4V voltage drop = big problem.

For 100A currents good connection of cables is essential. Switches must be rated at least for 100A. Using twisted pair will lower ElectroMagnetic noise generated.

I think, switching on and off Your oven, and measuring voltage at inverter input terminals at the same time, You can observe difference between on state and off state, showing voltage drop on cables between 12V source and inverter . Your multimeter should be stable and accurate enough for this purpose.

Do I understand your situation right: Your inverter is at the opposite end of the boat from your battery? If so, why not reduce most of your voltage drop problems by locating the inverter close to the battery and running the output from the inverter to the loads? This will reduce your current and voltage drop by a factor of 20, and place the noisey inverter a long way from your tv and appliances?

With properly insulated wire, and carefully following proper installation and grounding techniques, you should have no problems with safety issues.

HTRN

Hear hear HTRN..... isn't that why all electricity supply companies use high voltage distribution?

John.

Precisely. If he had stated from the beginning what his problem or objective was, it would have saved a lot of effort by everyone, and could have saved himself a lot of wiring cost.

Only after reading thru all the posts did I realize what he was trying to do. Afterwards, the solution was simple.

Court dismissed.

Call your next case.

HTRN

Rewiring my boat is not the problem. Nor is TV interference. I know exactly what to do about it.

My question is solely to do with the trust you can place in meters and how to check them against a voltage reference using a DIY approach.

I explained the boat situation so that readers would appreciate the importance of making decisions based on millivolt differences in 12vdc circuits when the meter reading itself was in doubt.

The question of the degree of precision and accuracy only arose because of the dozens of potential solutions depending on the reliability and trust required - and no doubt wanting to create a proper quality audit trail via certified test houses.

The (easiest) answer I liked (from Haajee) was to use the computer power supply as a reference source. But thanks for your contribution.

Sometimes we can't see the forest for the trees, and of that I guess I am guilty.

The answer was right in front of my nose, literally.

I like his suggestion of using computer power supplies as a reference.

Very obvious, simple, available, and easy.

As usual, we learn as well as teach on this forum.

I also send my thanks and regards to Haajee.

HTRN

Thanks my dear.

The accuray of measurement will only be confirmed if the program Speedfan be installed & readings of Multimeter are compared to indicated values.

A Tip for PC users & Service people:

Some problems of PCs eg rebooting without known reason are claimed by IT Support-people are due to decreasing outputs of PSU.

SpeedFan gives all power outputs indications well within better than 0.5 %

Regards

My reply to Forum

1.) If you need certaity of your measurements in your Service Job in repairing:

A good multimeter you have bought from market [has not been repaired yet] is good enough for your mental assurance. To be more assured you can buy another one , the same or a later & better one to compare when-ever you get confused. You need not to worry about CALIBRATION.

2.) If you are running a Repair Shop for Test-Equipment soley, then you need some CALIBRATION Facillity. But mind that the Calibrator sold by BIG_SHOTS like Fluke & many more itself need RECALIBRATION periodically. I had been working for 15 years in such set-up & have a Hard experience of all this. Vendors like Farnell & RS sell Test Equip in 2 ways, W/O Cal Service & with Cal service. But If you need Cal after repairs then you need your own facillity.

3.) There had been quite vast information on CV-Sources, Characeristics of it, & lot of new Tips & Know-how this discussion has imparted to. I thank all the subscribers.

Now I am offering my source of calibration:

1.) We all have PC

2.) At least one Power-Output-connector of PSU spare.

3.) It has 5V & 12V outputs.

4.) Your PC has quite accurate Sensors installed in Mother-board, CPU, HDD etc to measure Temperature, Supply Voltages CPU CoreV 3.3V, 5V, 12V & -12V & more but 5V & 12V are accesible for meter readings.

MB has quite accurate processing-circuitry to indicate Temp, Fan speed, PSU Voltages etc.

Chips used in MBs & HDDs etc for :

• Winbond:

W83781d, W83782d, W83783s, W83627HF, W83627EHF, W83627THF, W83637HF, W83697HF, W83L784, W83L785

• Via: Via 686, VT1211

• SMSC: LPC47Mxxx • ITE: ITE8705F, ITE8712F

• ASUS: AS99127F, AS99127F rev.2, ASB100

• National Semiconductors:

LM63, LM75, LM77, LM78, LM78-J, LM79, LM80, LM85, LM87, LM90, PC83630, PC83633, PC83634, PC83635, PC83636

• Analog Devices: ADT7476, ADM1027, ADM1028, ADM1030, ADM1031

• Maxim: MAX1617, MAX1618, MAX1619

• Genesys Logic: GL520, GL523 & more

5.) There are quite a number of SW [free] which indicate these readings in addition to info about PC SW & HW.

Some of those are:

a.) SpeedFan

This program supports:

Most Winbond sensors and ASUS AS99127F support fan speed changing, as well as others from MYSON, ANALOG DEVICES, NATIONAL SEMICONDUCTORS and ITE, but the motherboard should make use of available pins.

Lot of info on its Website about Sensors

Multi-Tab Indications like "Readings" indicates FanSpeed, Teperatures of all zones & PSU Voltages. & more

<http://www.almico.com/speedfan.php>

Author: Alfredo Milani-Comparetti <http://www.almico.com>

http://files3.majorgeeks.com/files/feb5e1c0c06a4737a6896b65f99808cd/diagnostics/installspeedfan431.exe

b.) Another program is in my PC @ my Home. It displays in a single Tab all the info about PSU O/Ps. I may post it later.

Thanks Haajee. You have given an enormous amount of info and it must have taken you much time to prepare your reply.

The one I like is the computer power supply - it should not be too difficult to do a quick check.

I will let you know how I get on.

Thanks & regards for all who read & replied to my post