### Author Topic: Using 9V batteries in series to calibrate a 6.5 digit multimeter  (Read 9460 times)

0 Members and 1 Guest are viewing this topic.

#### SharpEars

• Supporter
• Posts: 628
• Country:
##### Using 9V batteries in series to calibrate a 6.5 digit multimeter
« on: October 06, 2014, 01:22:30 pm »
I have a 6.5 digit multimeter that is calibrated pretty well up to 19.99999V. However in the 0-199.9999V range it is not faring very well. So, I was thinking.

I don't have a voltage standard I can use to calibrate the 199.9999V range. For this, I would need a precise voltage that is at least 60% of the range (i.e., at least 120V dc). Getting an accurate (transfer) voltage reference at this level is quite expensive. So, an idea struck me. What if I take 14 9v batteries and measure them in pairs using my multimeter's 19.99999V setting. I sum up the voltages (say 129.5V total if each battery nominally measures 9.25V on the average) and use that to calibrate my 199.9999V range. Since I know that the measurement of each battery pair will be very accurate (say to 20-30 ppm), I can foretell that my final calibration would be accurate to 140-210 ppm (i.e. 7 times 20-30 ppm) at the 129.5V (sum of all batteries in series) level. This level of accuracy (i.e., 140-210 ppm) is OK with me, since most of the high voltage measurements I need to take will be more in the 20-60V range at which point I will have much lower errors due to it being closer to the voltage of a single battery pair (which has 20-30 ppm error).

Am I correct in my assessment or is there some fundamental flaw to this scheme?
« Last Edit: October 06, 2014, 01:26:08 pm by SharpEars »

#### Gallymimus

• Regular Contributor
• Posts: 177
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #1 on: October 06, 2014, 01:30:37 pm »
Not the best idea, but not a terrible one considering lack of resources.  The best thing to do would be to string up the batteries, and make a reference measurement with a known good and accurate meter.  Can you borrow one from somebody?

If you do it your way, make sure you keep battery temps very consistent (i.e. don't warm them up in your hand too much) that will probably be your greatest source of error between when you measure each batt and try to measure the string.

#### SharpEars

• Supporter
• Posts: 628
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #2 on: October 06, 2014, 01:47:19 pm »
Not the best idea, but not a terrible one considering lack of resources.  The best thing to do would be to string up the batteries, and make a reference measurement with a known good and accurate meter.  Can you borrow one from somebody?

If you do it your way, make sure you keep battery temps very consistent (i.e. don't warm them up in your hand too much) that will probably be your greatest source of error between when you measure each batt and try to measure the string.

After connecting them in series, I will definitely let them settle to room temperature. I also thought about handling them with thick gloves on so as to avoid heating them up from contact with my skin. I understand the role that a steady temperature plays in this, so I will do my best to ensure that.

I wish I knew someone in the Chicagoland area who has a calibrated 7.5 or 8.5 digit meter and wouldn't mind me calibrating my Fluke 8506A from their meter. Alas, I don't...
« Last Edit: October 06, 2014, 01:49:52 pm by SharpEars »

#### rs20

• Super Contributor
• Posts: 2193
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #3 on: October 06, 2014, 01:57:30 pm »
I have a 6.5 digit multimeter that is calibrated pretty well up to 19.99999V. However in the 0-199.9999V range it is not faring very well. So, I was thinking.

I don't have a voltage standard I can use to calibrate the 199.9999V range. For this, I would need a precise voltage that is at least 60% of the range (i.e., at least 120V dc). Getting an accurate (transfer) voltage reference at this level is quite expensive. So, an idea struck me. What if I take 14 9v batteries and measure them in pairs using my multimeter's 19.99999V setting. I sum up the voltages (say 129.5V total if each battery nominally measures 9.25V on the average) and use that to calibrate my 199.9999V range. Since I know that the measurement of each battery pair will be very accurate (say to 20-30 ppm), I can foretell that my final calibration would be accurate to 140-210 ppm (i.e. 7 times 20-30 ppm) at the 129.5V (sum of all batteries in series) level. This level of accuracy (i.e., 140-210 ppm) is OK with me, since most of the high voltage measurements I need to take will be more in the 20-60V range at which point I will have much lower errors due to it being closer to the voltage of a single battery pair (which has 20-30 ppm error).

Am I correct in my assessment or is there some fundamental flaw to this scheme?

Fwiw, the ppm accuracy doesn't mount up like that. If you have 7 batteries measured to 10 ppm, then you have the sum of those batteries to 10 ppm too. If it's not clear why, try substituting real numbers:

9V +/- 10 ppm is 9V +/- 90 uV
7 in series is 63V +/- 630uV, which is 63V +/- 10 ppm.

(If the individual errors were uncorrelated, the ppm total error would actually be smaller than the individual readings! Just like averaging noisy data to get rid of the noise) So, assuming that batteries are ideal voltage sources, your idea is quite good. Whether the batteries would show the same voltages under the different currents imposed on the batteries in the two different ranges is another question, probably rendering my above point irrelevant  .

#### wiss

• Frequent Contributor
• Posts: 485
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #4 on: October 06, 2014, 02:01:43 pm »
You could build a resistive divider, say 8:1, and verify it at 18 V in (measure 18 V and ~2.2 V on the 19 V range), then increase the input voltage to 144 V (199 V range) and compare to the divided 18 V (19 V range).

#### Fungus

• Super Contributor
• Posts: 11094
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #5 on: October 06, 2014, 02:15:45 pm »
After connecting them in series, I will definitely let them settle to room temperature. I also thought about handling them with thick gloves on so as to avoid heating them up from contact with my skin.

Why would you need to touch them once they're all connected up? You can still measure the pairs when they're all in series.

Try to spend as little time as possible with the multimeter connected.

It might be worth adding some capacitors into the mix. Something low leakage (ceramic) across each battery and another big one across the whole array. Just because of cosmic rays, etc.

Apart from that ... it sounds like it might work.

#### wiss

• Frequent Contributor
• Posts: 485
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #6 on: October 06, 2014, 02:19:47 pm »
Maybe we should add a warning here?

#### SharpEars

• Supporter
• Posts: 628
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #7 on: October 06, 2014, 02:40:22 pm »
Fwiw, the ppm accuracy doesn't mount up like that. If you have 7 batteries measured to 10 ppm, then you have the sum of those batteries to 10 ppm too. If it's not clear why, try substituting real numbers:

9V +/- 10 ppm is 9V +/- 90 uV
7 in series is 63V +/- 630uV, which is 63V +/- 10 ppm.

(If the individual errors were uncorrelated, the ppm total error would actually be smaller than the individual readings! Just like averaging noisy data to get rid of the noise) So, assuming that batteries are ideal voltage sources, your idea is quite good. Whether the batteries would show the same voltages under the different currents imposed on the batteries in the two different ranges is another question, probably rendering my above point irrelevant  .

Sadly the error in the individual readings must be at least partially correlated. Given a 20-30 ppm error, some of it is random and some of it is due to the MM reading over/under correct voltage. The over/under correct voltage part will accumulate over multiple readings.
« Last Edit: October 06, 2014, 02:53:42 pm by SharpEars »

#### MarkPalmer

• Regular Contributor
• Posts: 101
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #8 on: October 06, 2014, 02:59:40 pm »
Watch your full scale ratings- The 8506A has 100mV, 1V, 10V, 100V, and 1kV DCV ranges.  I’ll take it you have the first three fairly good.  Full scale on the 100 VDC range is 128 V per the manual.  I believe you only need 60% of the 100 VDC of the scale, or an accurate source of 60 volts or more to software calibrate it.  (Fewer 9 V batteries.)  600+ volts would be needed for the 1 kV range.

-Mark-

#### SharpEars

• Supporter
• Posts: 628
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #9 on: October 06, 2014, 03:54:37 pm »
Watch your full scale ratings- The 8506A has 100mV, 1V, 10V, 100V, and 1kV DCV ranges.  I’ll take it you have the first three fairly good.  Full scale on the 100 VDC range is 128 V per the manual.  I believe you only need 60% of the 100 VDC of the scale, or an accurate source of 60 volts or more to software calibrate it.  (Fewer 9 V batteries.)  600+ volts would be needed for the 1 kV range.

-Mark-

Thanks Mark, good catch on the 128V max. I guess I would need 12-13 batteries to get up to 95% full scale and not saturate the ADC, per the calibration warnings.

#### David Hess

• Super Contributor
• Posts: 11662
• Country:
• DavidH
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #10 on: October 07, 2014, 01:00:27 am »
It would be better to do something like this:

Stacking the REF50xx for High-Voltage References

http://www.ti.com/lit/an/sbaa203/sbaa203.pdf

#### DanielS

• Frequent Contributor
• Posts: 798
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #11 on: October 07, 2014, 02:08:42 am »
It would be better to do something like this:

Stacking the REF50xx for High-Voltage References

http://www.ti.com/lit/an/sbaa203/sbaa203.pdf
That's one crazy use of a low-voltage reference IC!

Better hope nothing ever fails open on the 100kV version.

#### kwass

• Frequent Contributor
• Posts: 322
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #12 on: October 07, 2014, 02:37:06 am »
.....
Am I correct in my assessment or is there some fundamental flaw to this scheme?

With 6.5 digits you'll see the self-discharge of a 9 volt battery.  Assume a 10 year self discharge to zero volts that would be about 0.000002 volts/minute.  So even if you carefully control the temperature, you'll have to get this measurement accomplished pretty fast to eliminate that source of error.
-katie

#### zapta

• Super Contributor
• Posts: 6004
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #13 on: October 07, 2014, 03:10:42 am »
Does the higher range have proportionally higher resistance? Otherwise the battery load will be different in the single and overall measurements.
Drain the swamp.

#### David Hess

• Super Contributor
• Posts: 11662
• Country:
• DavidH
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #14 on: October 07, 2014, 03:32:06 am »
Does the higher range have proportionally higher resistance? Otherwise the battery load will be different in the single and overall measurements.

For meters which support it, the low ranges will be the ones with the high input resistance.  The high voltage ranges will use a 10 Mohm input resistance divider.  Electrometers have high input resistance on all ranges.

In this case it is a Fluke 8506A so the 10 volt and below DC ranges have high input resistance and the 100 volt and 1000 volt ranges have a 10 Mohm input resistance.  That is better than any of my meters which only have a high input resistance up to 2 or 3 volts.

• Regular Contributor
• Posts: 212
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #15 on: October 07, 2014, 03:59:27 am »
Has anyone done this?

It would be better to do something like this:

Stacking the REF50xx for High-Voltage References

http://www.ti.com/lit/an/sbaa203/sbaa203.pdf
That's one crazy use of a low-voltage reference IC!

Better hope nothing ever fails open on the 100kV version.

#### SeanB

• Super Contributor
• Posts: 15408
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #16 on: October 07, 2014, 04:38:53 am »
I probably would go and buy a whole box or three of CR1610 cells and stack them in a plastic tube. 40 will give a more stable 120V or so transfer standard. Lithium cells have a much lower TC and a more stable voltage. You could stack in your application into small packs and measure the voltage then simply put in series for the final test. Just handle them with cotton gloves to keep from getting them contaminated and they should hold stable for a short term long enough to do the calibration.

#### David Hess

• Super Contributor
• Posts: 11662
• Country:
• DavidH
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #17 on: October 07, 2014, 04:39:49 am »
Has anyone done this?

The application note includes photos of the 100 volt strips, a 1000 volt module, and the finished 100 kilovolt assembly so I assume they did it.

Operating 3 terminal series references as 2 terminal shunt references and stacking them is not new but these guys took it to a whole different level.  This is a great way to build a high voltage reference which may be calibrated against a lower voltage one.

#### David Hess

• Super Contributor
• Posts: 11662
• Country:
• DavidH
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #18 on: October 07, 2014, 04:51:21 am »
I probably would go and buy a whole box or three of CR1610 cells and stack them in a plastic tube. 40 will give a more stable 120V or so transfer standard. Lithium cells have a much lower TC and a more stable voltage. You could stack in your application into small packs and measure the voltage then simply put in series for the final test. Just handle them with cotton gloves to keep from getting them contaminated and they should hold stable for a short term long enough to do the calibration.

They used to use mercury cells (1.35 volts) and then later silver oxide cells (1.55 volts) as voltage references.  I remember old voltmeter kits which used a mercury cell for calibration.

#### NANDBlog

• Super Contributor
• Posts: 4733
• Country:
• Current job: ATEX certified product design
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #19 on: October 16, 2014, 11:43:49 am »
With a high voltage amplifier like the LT6090, accurate voltage dividers (bulk metal foil), and an accurate voltage reference, I believe it is more than possible to build a high reference voltage with 30-50ppm accuracy.

#### macboy

• Super Contributor
• Posts: 2019
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #20 on: October 16, 2014, 11:58:11 am »
.....
Am I correct in my assessment or is there some fundamental flaw to this scheme?

With 6.5 digits you'll see the self-discharge of a 9 volt battery.  Assume a 10 year self discharge to zero volts that would be about 0.000002 volts/minute.  So even if you carefully control the temperature, you'll have to get this measurement accomplished pretty fast to eliminate that source of error.
You definitely can't assume a 10 year slope to zero. At 10 years, an alkaline cell will still measure at it nominal voltage and will be usable with a significant percentage of its nominal capacity left. I have some AA alkaline cells that I pulled out of a box recently; I know them to be at least 20 years old, and they still measure just over 1.5 V each. Too bad they don't have a date code on them, I'd like to know just how old they really are.

#### kwass

• Frequent Contributor
• Posts: 322
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #21 on: October 16, 2014, 02:59:40 pm »

You definitely can't assume a 10 year slope to zero. At 10 years, an alkaline cell will still measure at it nominal voltage and will be usable with a significant percentage of its nominal capacity left. I have some AA alkaline cells that I pulled out of a box recently; I know them to be at least 20 years old, and they still measure just over 1.5 V each. Too bad they don't have a date code on them, I'd like to know just how old they really are.

I know that you can't assume a linear discharge, especially at no load, but I could find a spec for this.    I was doubling the specified shelf life as found on the Energizer spec sheet for AAAA cells:

http://data.energizer.com/PDFs/E96.pdf

AAAA cells (6 of them in a 9 volt battery) are particularly bad at self discharge compared to AA cells.  Energizer says the AA has a 10 year shelf life while only 5 years for the AAAA.
-katie

#### Galaxyrise

• Frequent Contributor
• Posts: 527
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #22 on: October 16, 2014, 04:34:18 pm »
I went down a similar line of thought, stacking up 9V batteries and measuring the voltage.  As has been mentioned above, I could readily see the voltage changing.  I did have lower voltage references, so I used a divider and an opamp to get 100V as stable as my low voltage reference and the voltage divider:
https://www.eevblog.com/forum/beginners/why-does-my-op-amp-circuit-need-a-capacitor/

The circuit I list at the end is missing a couple components, most notably a current limiting resistor.  But it worked pretty well!
I am but an egg

#### babysitter

• Frequent Contributor
• Posts: 848
• Country:
• pushing silicon at work
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #23 on: October 16, 2014, 08:28:43 pm »
Why not touching the batteries with your warm hands?

Really attach them close to your body with insulation against the anbient- thats one of the best thermostats you can get for 0 dollar!

Use always the same body parts. Orifices might work better than surface, torso better than periphery. Best performance should be at rest at the same time of the day.

I'm not a feature, I'm a bug! ARC DG3HDA

#### wiss

• Frequent Contributor
• Posts: 485
• Country:
##### Re: Using 9V batteries in series to calibrate a 6.5 digit multimeter
« Reply #24 on: October 16, 2014, 10:04:05 pm »
Why not touching the batteries with your warm hands?

Really attach them close to your body with insulation against the anbient- thats one of the best thermostats you can get for 0 dollar!

Use always the same body parts. Orifices might work better than surface, torso better than periphery. Best performance should be at rest at the same time of the day.

Start out with a AAA and eventually uppgrade to  a D for increased stabillity, might require some practice...

Smf