Burette classes and measurement

[Bemused]

Hi all,

Class A burettes (50ml) have a stated tolerance of +/- 0.05ml while class B are +/- 0.1ml.  What are the practical implications of this in recording measurements? 

Typical practice is to estimate a second decimal place which seems wrong with class B burettes which are most commonly in use in schools - a titre would be recorded as e.g. 25.55 +/- 0.1 ml.  Here the number of significant digits are outside the tolerance or is the actual reading to be taken as somewhere between 25.45 and 25.65?  In the case of a class A then this would be between 25.50 and 25.60.  In essence, it seems wrong to record more significant digits than the tolerance allows - this would be rejected in engineering.

I'd appreciate thoughts, thanks.

[studiot]

2 hours ago, Bemused said:

Hi all,

Class A burettes (50ml) have a stated tolerance of +/- 0.05ml while class B are +/- 0.1ml.  What are the practical implications of this in recording measurements? 

Typical practice is to estimate a second decimal place which seems wrong with class B burettes which are most commonly in use in schools - a titre would be recorded as e.g. 25.55 +/- 0.1 ml.  Here the number of significant digits are outside the tolerance or is the actual reading to be taken as somewhere between 25.45 and 25.65?  In the case of a class A then this would be between 25.50 and 25.60.  In essence, it seems wrong to record more significant digits than the tolerance allows - this would be rejected in engineering.

I'd appreciate thoughts, thanks.

You need to distinguish between tolerance (which affects accuracy) and reading resolution (which affects precision).

Tolerance is a specification of full scale ie a nominal delivery of 50ml will be within the range 49.95mL to 50.05 mL.
This is a characteristic of the instrument (in this case a burette or buret in US)
This gives a relative tolerance of 0.13%

The actual scale can be read more finely than this.

 

This effect is common in many instruments where a scale is read rather than a digital readout obtained.

For example mechanical verniers, analogue voltmeters.

For both of these latter tolerance is usually given as the relative (of full scale) %.
Such a voltmeter would always be specified as 1% (very good) 5% (good) 10%(El cheapo variety).

Remember for readings that are made by difference (as in a burette) the tolerance is a systematic error that affects both reading equally and with the same sign so cancels out on subtraction.

That leaves reading resolution.

This is measured as the smallest scale graduated interval (usually 0.1mL for a 50 mL burette), or some fraction of it.

(ASTM E287-02) has half of this at 0.05mL.

Vogel suggests this for "all ordinary work", but reading to 0.01 or 0.02 mL with the aid of a lens, "for precision work".

Precision work could also entail establishing a calibration curve for use over the whole length of the burette or alternatively repeating the measurement over several different parts of the sscale and averaging.

 

 

Edited August 18, 2019 by studiot

[Bemused]

Thanks Studiot,

So the measurement ranges I gave in my example for a Class A or B burette are correct then?  Or does the printed tolerance on a burette only refer to the full scale nominal delivery you mention above (going from 0 to 50)?  Essentially what I am asking is how would you report a titre result for a class A (+/-0.05 ml) and class B (+/-0.1ml) burette.

 

 

[Bemused]

Actually, thinking about this some more, the tolerance must refer to the delivery of the total volume (50ml), right?  Otherwise delivery of small volumes, say 0.1ml would have 100% uncertainty!  If the tolerance applies over the delivery of the full volume then the uncertainty per division is 0.1/500 = 0.0002ml for a class B and 0.05/500=0.0001ml for Class A.   

[studiot]

1 hour ago, Bemused said:

Actually, thinking about this some more, the tolerance must refer to the delivery of the total volume (50ml), right?  Otherwise delivery of small volumes, say 0.1ml would have 100% uncertainty!  If the tolerance applies over the delivery of the full volume then the uncertainty per division is 0.1/500 = 0.0002ml for a class B and 0.05/500=0.0001ml for Class A.   

You are getting there.

The tolerance standards refer to laboratory glassware in general and include glassware (flasks, pipettes etc) that have only one measuring line and no scale.
As such they are designed to deliver the rated quantity /- rated tolerance.

Since you compared this with engineering practice, consider this.

A tape measure with one inch missing from the end will 'measure' a 10 foot wall as 10 feet and 1 inch, from the end.

But by measuring from the 1foot marking to the 11 foot marking, the correct length of 10 feet will be obtained.

This is because the tape has an error everywhere of 1 inch so using a mesurement by difference this cancels.

Length of wall = (11' 0") - (1' 0") = 10 feet.

The burette allows this measurement by difference so the resolution and accuracy are not necessarily limited by the overall tolerance.

But inEngineering you also have 'limits and fits' tolerance.

Suppose you are turning a bar down to 'just fit' through a particular hole.
What tolerance would you enforce to turning the bar down?
This is equivalent to using a measuring flask where you cannot measure volume by difference.
Your tolerance refers to the whole diameter and again you cannot use difference.

Remember also that there are other considerations that affect accuracy with a burette and if you try to get too accurate you need to start to consider air bubbles, temperature, liquid density and the scale accuracy of the burette and techniques to ensure that all the liquid you think is transferred actually reaches the receiving vessel.
At this point you might move to weight measurement rather than volume.

Edited August 19, 2019 by studiot
correction to calculation.

[Bemused]

Does your analogy with a measuring tape not break down though in that the "error" is only one-way and fixed?  Each burette measurement could be on either side of the true measurement and to a different degree.  Hence the tolerance can influence the measurement.

Also, thinking about glassware with just a single mark, you mentioned in your initial post that tolerance affects accuracy while reading resolution affects precision.  Taking the case of a pipette though (say 25ml and +/-0.05ml) you would report the measurement as 25.00 +/- 0.05ml right?  In this case the tolerance also defines or at least bounds the precision, no?

Finally, can you confirm that the tolerance listed on a burette refers to the delivery of the total volume e.g. 50ml +/- 0.1ml (for Class B) and not to each individual delivery e.g. 1ml +/- 0.1ml?

Appreciate your help with this. 

[studiot]

On 8/20/2019 at 12:49 PM, Bemused said:

Does your analogy with a measuring tape not break down though in that the "error" is only one-way and fixed?  Each burette measurement could be on either side of the true measurement and to a different degree.  Hence the tolerance can influence the measurement.

Also, thinking about glassware with just a single mark, you mentioned in your initial post that tolerance affects accuracy while reading resolution affects precision.  Taking the case of a pipette though (say 25ml and +/-0.05ml) you would report the measurement as 25.00 +/- 0.05ml right?  In this case the tolerance also defines or at least bounds the precision, no?

1) Finally, can you confirm that the tolerance listed on a burette refers to the delivery of the total volume e.g. 50ml +/- 0.1ml (for Class B) and not to each individual delivery e.g. 1ml +/- 0.1ml?

Appreciate your help with this. 

(1) Yes, see this extract from the latest version of  'Pharmaceutical Analysis' by Watson.

Note the tolerance on the entry for the burette.

 

 

As a matter of interest what tolerance would you ascribe when the buretter reads 0.00?

Would you allow a negative volume?