Voltage madness

[nathanlee]

I've just sent this email to Nasa marine, but while I'm waiting for a reply I thought I'd see what you clever folks have to say.

---------------------
Last year I bought one of your excellent BM-1 battery monitors, and although I must confess that without an engine or a connection to shore power for a week, the fickle charge from my solar cells made the remaining capacity "percentage" reading a bit useless. Not to worry, I'm not interested in that number anyway. The ability to see current going in and out of the battery is the reason I bought it.

I've recently totally rebuilt my boat's electrical system. Every single component has been replaced, from switch panel to battery, and I've stumbled across a strange problem. The BM-1 seems to be over reading the voltage of the battery. I first noticed this when testing for losses in my cable runs, and upon finding unexpected results, I put the multi meter across the terminals of my 200ah AGM battery, and found 13.03v, where as the BM-1 was reading 13.3v.

I've since borrowed another, more expensive multimeter, and tried again. Here are the results for the particular state of charge the battery is in now. The values in brackets are the difference.
BM-1: 12.9v
cheap meter: 12.73 (0.17)
expensive meter: 12.87 (.03)

Now, flicking the shore power charger on, bumps the voltage up to

BM-1: 14.4
cheap meter: 14.08 (.32)
expensive meter: 14.25 (.15)

---------------------

Actually, one thing I've noticed since sending the email, is that the Sterling charger is set to AGM mode, which I've just read in the manual, gives a charge voltage of 14.4, suggesting the BM-1 is correct. However, I've no way to prove this given the difference in all three readings. I'm confused.

Bloody electrons!

Cheers,

Nathan

 

[malin9897]

are the different readings not just down to the calibration of the different meters

 

[peteK]

Nathan,the one thing I noticed with cheap meters is they lose their accuracy sooner than the expensive meter.If you take the back off there is usually an adjustment to reclibrate the meter.Use the expensive meter to compare with ,you will probably find it wont match on all voltages.The BM1 and expensive meter voltages are near enough the same as not to matter.

 

[lenseman]

Firstly, you have discovered a very common 'fault' with multimeters, if you use twenty DVM's (Digital Volt Meters) to do the same measurement, you will possibly get twenty different answers. If you were using an analogue volt meter, you probably would not really notice the reading error?

That said, the meters if taken independantly, can be a very useful tool only if you understand that the spot measurement is technically a relative measurement that can ony be used as a 'reading' if you use the same meter to take the second and subsequent measurements.

Regarding the BM-1, you possibly might find a 'skeli-pot' (variable resistor) inside which can be adjusted to change and match the BM-1 meter reading to that of other DVM test meters. See notes below before adjusting any meter.

You would than have all meters reading the same. This is often called 'transfer calibration' as long as you can adjust the meter with the incorrect reading to one that you know is correct.

First, you need to find an electronics engineer who has a UKAS (United Kingdom Accreditation Service) calibrated volt meter which has been compared to the NPL standard (National Physical Laboratory) and is within the calibration date! Your DVM can then be checked against this Standard and any errors written down for you to use as an off-set which you can then add or subtract when you take your own measurements on Kudu's electrical circuits.

I hope this helps?

 

[fergie_mac66]

assume you have zeroed the meter and alLowed for parallax using the mirror scale

edit : well he said it now !! mirrored scale specialy usefull on digi meters !

 

[EuanMcKenzie]

so what

seriously its not a problem as long as you use the same meter all the time as you trying to detect change and I suspect you will find less errors when monitoring drop in voltage.

Are you really going to run your batteries right down anyway.

you understand its a calibration difference.

Personally I would go and find another harder job to finish.

 

[savageseadog]

I can't explain your meter readings off charge other than to suggest that with meters you get what you pay for. I have come across a drastically poor meter which wasn't a dirt cheap one fairly recently.
Measuring voltages on charge can lead to misleading readings. A small AC voltage of not necessarily regular waveform will be impressed onto a steady DC voltage. What various meters will actually read will depend on the analogue to digital conversion technique. A professional meter like a Fluke will have a full specification from which you can work out what's happening.

 

[nathanlee]

I've got to be honest, I was sort of hoping for an answer like "your fluffle valve is sticking".

Just goes to show that while I can wire a plug, and if really pressed, fudge a pcb together, I actually have no idea about electronics.

So I guess the practical answer is; my meter is borked, the good meter is accurate enough. use that one, jobs a goodun?

I understand that I very accurate reading is probably required if you're working out the swing for some precise rf amplifier, but what's a sensible tolerance for generic on board 12v testing?

And thank you all.

 

[concentrik]

BM-1: 12.9v
cheap meter: 12.73 (0.17)
expensive meter: 12.87 (.03)

Now, flicking the shore power charger on, bumps the voltage up to

BM-1: 14.4
cheap meter: 14.08 (.32)
expensive meter: 14.25 (.15)

---------------------

Nathan

Hello Nathan,

Others have said it, but it'll just be calibration (or conversion, analogue to digital) error which is maddening but normal in different instruments. That said, they each may be within their specified tolerance - the differences you note are under 3 percent (I think, check my maths!)

I don't think you can reliably measure the charger output with a meter. As has been said it could be there's a 'ripple' component to the charger output, which different instruments would interpret in their own unique (different) way. Ideally you'd need an oscilloscope to look at the ripple component. The charger may even be delivering a pulsed/ramped output although I don't know how sophisticted that unit is.

In he end the differences are real but probably of no consequence.

Andy

 

[westernman]

Many moons ago I worked in the quality assurance department of a company which made industrial measurement equipment. My job was to recalibrate all the measurement equipment used on the production line according to its schedule and have a complete calibration trail going back to the physical standards at the NPL.

From what I remember, the upmarket voltmeters calibration requirement was +/- 2% and they had to be rechecked every 6 months. I would not expect anything you or I are likely to buy at the local hardware store to get anywhere close to that good.

 

[concentrik]

and another thing....

sorry I didn't see your last post....

I think you're looking for a reliable indication of state-of-charge rather than accuracy. Battery charge/discharge rates and capacity rely on a number of external factors, not least (wait for it....) temperature. I expect whatever meter you use you'll get to know the system very well, recognising that a particular reading means batteries are as charged as they usually are, and lower readings take you nearer to a terminal voltage where things start to go to sleep.... VHF won't work below such-and-such a reading, tiller pilot gives up at another. It's repeatability and "internal accuracy" you need rather than absolute accuracy. As such the BM1 is probably as good as anything!

Andy

 

[elton]

By their nature, digital readouts give the impression they are more precise than they actually are. But they generally are sufficiently precise for all practical purposes. The difference you're seeing of a few percent is as likely to be caused by different probes and leads, or different points of contact of the probes, as differences in accuracy of the instruments.

 

[VicS]

By their nature, digital readouts give the impression they are more precise than they actually are. But they generally are sufficiently precise for all practical purposes.

What you meant to say is:
"By their nature, digital readouts give the impression they are more acccurate than they actually are. But they generally are sufficiently accurate for all practical purposes.

The trouble stems from the fact that you can get a precise reading, usually more precise than from analogue meter, but that does not mean they are more accurate.

 

[elton]

What you meant to say is:
"By their nature, digital readouts give the impression they are more acccurate than they actually are. But they generally are sufficiently accurate for all practical purposes.

The trouble stems from the fact that you can get a precise reading, usually more precise than from analogue meter, but that does not mean they are more accurate.

True enough. You might say I was somewhat sloppy in my terminology, but I think 'precise' would be the layman's term of choice in this context. I can sense a debate coming on! It's the degree of precision that leads to an unrealistic expectation of accuracy, put it that way

 

[noelex]

As others have said multimeters are often poorly calibrated. It is worth trying to measure the error in them from time to time so you have a better idea of the correct reading.
I use a special, expensive, lithium battery charger which has a very precise (I hope) voltage output to check the calibration of mine.
I do remember reading a method of measuring the voltage of a new alkaline 9v battery. The correct reading should be 9.??. The web site claimed the voltage of different brands etc was quite consistent and a multimeter could be calibrated reasonably accurately by noting the error.
Unfortunately I cannot find the website. Does anyone know it? or a similar easy method?
Given the poor calibration of multimeters something that gave a reference that was +/- 0.1v or so would be an improvement.

 

[lenseman]

. . . . .
BM-1: 12.9v
cheap meter: 12.73 (0.17)
expensive meter: 12.87 (.03)

Now, flicking the shore power charger on, bumps the voltage up to

BM-1: 14.4
cheap meter: 14.08 (.32)
expensive meter: 14.25 (.15)

. . . . . .

Forget the readings in your brakets, look at the difference in reading another way.

EUT . . . . . . . . . Shore power charger Off . . . Shore power charger On . .. .CHANGE

BM-1 . . . . . . . . . . . . . . 12.9v . . . . . . . . . . . . . . . .14.4v. . . . . . . . . . . . . . 1.5v

Cheap meter. . . . . . . . . 12.73v . . . . . . . . . . . . . . 14.08v . . . . . . . . . . . .. .1.35v

Expensive meter . . . .. . 12.87v . . . . . . . . . . . . . . 14.25v . . . . . . . .. . . . . .1.38v

It is the change of voltages which need to be compared and as you can see, there is not a great deal of difference.

 

[charles_reed]

What you meant to say is:
"By their nature, digital readouts give the impression they are more acccurate than they actually are. But they generally are sufficiently accurate for all practical purposes.

The trouble stems from the fact that you can get a precise reading, usually more precise than from analogue meter, but that does not mean they are more accurate.

Now we start the famous precision vs accuracy discussion. It's all down to that 2nd decimal place.

Etymologically Elton is probably correct, thought the adjective "apparent" would be a valuable addition.

If the OP is complaining to NASA I think he's being unfair - the voltages are all within normal variations.

 

[Plevier]

You are reading on the battery terminals with the multimeter I think.
Is the BM1 reading there or from the bus?
If there was significant charging or load current when you took the readings, are you actually measuring the voltage drop in part of your wiring?

 

[savageseadog]

High resolution is the terminology, accuracy is another thing dependant on temperature, quality of the components, meter battery voltage and so on. Then there is the precision and method of the digitisation and how the input is prefiltered including the integration period, which will determine the response to different waveforms and their periods. Then there will be electrothermal factors, different metals at different temperatures. Then there is the measurement method.

You have a 4 digit meter 9or 3and a half). The last digit resolves .01V on the range you will be using. Mathematically that last digit can be considered plus or minus one to begin with so all of the above plus or minus one digit. If you can afford a quality meter like a Fluke it will last a lifetime and provide you with reliable quantifiable results.

 

[nathanlee]

You are reading on the battery terminals with the multimeter I think.
Is the BM1 reading there or from the bus?
If there was significant charging or load current when you took the readings, are you actually measuring the voltage drop in part of your wiring?

BM1 is reading from the shunt, which is a 2 inches away from the negative terminal, and 6 inches, a fuse, and a switch away from the isolator switch. Of course, then there's the 5 core cable, running about a meter to the BM-1 unit.

If anything, the BM1 should be under reading do the voltage drop, not the other way around.

It's ok, I'm happy with my results now. I was splitting hair expecting anything more accurate, it's just that I'd never seen the phenomenon before since I'd never looked at two independent volt meters on the same bit of kit. As as been suggested, I just presumed that with it being a digital multimeter, it'd be perfectly accurate.

Thanks again folks.