Over the years, customers of Primary Medical have asked about the relationship between our calibration during preventive maintenance and their field tests. This note addresses those differences, the most common reasons for apparent calibration errors, and some thoughts on procedures you might want to use to save time and cost while improving your field operations. There are fundamental differences between the instruments used in a hospital environment and those we employ to calibrate a vaporizer in our facility. Your operations require portability, durability, and simplicity of use; ours must be reliable, accurate, and traceable. In practice, field operations almost always rely on the Riken Model 18 Portable Gas Indicator, an admirable hand-held interferometer ideally suited to relative measurements of vaporizer performance in the field. Primary Medical's principal instrument is the Foxboro Miran 1A CVF General Purpose Laboratory Analyzer. It is a precise bench spectrophotometer which determines total gas content by differential absorption. The accuracy of an instrument is determined both by its inherent capability and human error. On both counts, the Miran on the test bench outperforms the Riken in the field
The Riken interferometer (above photo) measures the difference in effective path length between a reference (a sealed volume of air) and the test gas. The effective paths would be identical if the gases had the same density and index of refraction. The index of refraction depends on the composition of the gases, which is what we are trying to determine. But the density depends on the temperatures and pressure of the test gas and on the temperature of the reference. Riken estimates that under normal operating conditions the Model 18 provides accuracy in the order of 15% of the reading, or 3% of full scale. In contrast, the Miran chromatograph measures the total amount of the selected gas in the path. That measurement is accurate to .5%. (Temperature variations could increase the error significantly, except that our calibration is performed over a wide range of temperatures.)
The first source of human error is in reading the instrument: where the Riken provides broad gray fringes on a light-gray background, the Miran reads out directly on a digital display or a mirrored analog scale. The same advantage is seen in zero-setting, where the Riken again requires judgement about fringe locations and the Riken reads a distinct null. Correction of the indicated reading for anesthetic agent on the Riken is both critical and straightforward, but even the best technicians have occasionally failed to apply it correctly; due to the distinct difference in absorption bands for the Miran and the fact that no correction is required, there is virtually no chance that the choice of agent could be a source of error in its use.
Primary Medical determined early in its operation that while the Riken is acceptable in the field, it is not adequate for our calibration. If the Miran is so much more accurate than the Riken, should you be using it as we do? Surprisingly, the general answer is "No." It is quite expensive, much more difficult to use correctly, and suitable only for the laboratory bench, not for the surgical theater. In the operating room, you need speed and ease of use far more than precision. In the right environment, with highly trained personnel, disregarding cost, and providing for calibration of the chromatograph, the Foxboro is better; in a hospital, the Riken is usually right. We recommend that you check each vaporizer with your Riken when it arrives. At any setting of the flow dial, it should read within about 15% of that setting. If you find a consistent bias over the units you receive, please check the condition of your Riken. If one unit we return does not fit the pattern of the others, please check with us and, if necessary, send it back to us for verification. Primary Medical's policy is that there is no charge for re-calibration if a vaporizer is out of tolerance on retest; we will charge for shipping if we find no error on our part.
The following checklist may save both of us time and money in the event that your check suggests a calibration error.
1. Is the error within the allowance for the Riken (about .2% to 15% of reading)? If not, please retest and attempt to minimize field test errors before assuming the calibration to be at fault; usually, the fault is in the field measurement.
2. Are other well-calibrated vaporizers reading similarly? If so, there is likely to be a systematic error introduced in your operations: failure to set zero correctly, bias in reading the interference fringes, or improper purging, pressurization, or stabilization of the test gas.
3. Is there evidence of physical damage, especially in the packing materials? If so, the fault is likely to be either in your receiving procedures or in the shippers handling of the unit.
4. Have you checked with us? Please give us a call before returning a vaporizer, so that we can pull its records. We can determine from those records whether its operation was marginal; if we have serviced it more than once, we can evaluate whether major maintenance should be initiated. We will be especially grateful for such a call if it helps us identify any error in our operations.
We have often said "analysis alone is not enough but routine maintenance is." That applies to the test instruments as well as to the vaporizers themselves. Primary Medical is meticulous in maintaining our instrumentation as well as your vaporizers. The Riken is a fine instrument, but its inherent errors, imprecise reading, and failure of calibration can suggest errors in our maintenance which simply are not real. Even Primary Medical can't fix problems which don't exist.