Testing and analytical equipment selection and assessment

Testing and analytical instrumentation, while necessary for many businesses, comes at a high capital and maintenance cost. Getting this decision wrong can saddle a business with an expensive capital asset for years to come. We have seen this happen when the problem statement was inadequate. With 33 years experience in materials science and chemistry research, we have used a wide range of instruments including optical microscopes, electron microscopes, spectrometers, thermal analysis equipment, x-ray diffractometers, electrochemical instrumentation, mechanical test frames (tensile, fatigue, creep) and dimensional measurement equipment. The list is long and it needs to be made clear that in most cases our experience is: identifying a test capability; specifying the test; analyzing and reporting results. We are not the user or operator interested in the most state of the art capability, rather we are the customer interested in results to solve a problem. This puts us in a unique position to help your business with testing and analytical equipment selection. The process is generally as follows:

Identify the problem
Identify the types of instrumentation critical to solving your problem
Assess the need for instrumentation in-house vs. at outside laboratories
Prepare preliminary specifications
Identify makers of instrumentation potentially meeting the specification
Consult with instrument makers to short list suitable instruments

Case studies in adaptation of older instruments to new tasks for cost savings and new product development

We take a fundamental approach to instrumentation to understand the underlying physics of the measurement. As collectors and restorers of historic scientific instruments, we recognize that these instruments are often closer to the fundamentals and less of a “black-box”. Sometimes, an older used instrument or method can provide the information you need at significantly lower cost. For example, modern optical microscopes can break the budget. However, many research microscopes made in the second half of the 20th century, state of the art for their time, can provide near state of the art imaging when simple modifications are made to collect photographic images. The examples below show how mid 20th century equipment can still solve modern science and engineering problems. We are not “Luddites”, but realists when it comes to understanding the cost/benefit analysis of an instrument.

Spectroscopy

Spectrometers typically represent a large capital expense for an organization. Modern instruments are great for their quantitative accuracy and relative ease of use. However, there are circumstances where only a qualitative assessment of performance is required. If the spectral range of interest is in the visible spectrum, an inexpensive Bunsen spectroscope can be fitted with a modern digital camera to capture the spectrum. Due to the non-linear refractive index vs. wavelength effect on prism dispersion, such instruments represent a challenge for quantitative analysis. The following is based on our initial research into making low cost spectrometers designed for specific applications. It was inspired by a request from a home owner’s association to assess the lighting performance of sodium vapor vs. LED lamps for outdoor area illumination. However, while wavelength and quantitative intensity were not critical to answering the customer’s question, they are absolutely critical components of any spectrometer design intended for quantitative analysis.

The example shown at right was prepared to asses the spectral performance of a modern LED light bulb (top). The middle spectrum is from a low pressure sodium lamp. The prominent sodium D-doublet lines are evident in this spectrum, but so are other much less intense lines which are from impurities. Note that the sodium lamp spectrum extends well into the visible violet range, whereas the LED light source only reaches the indigo. The bottom spectrum is from a hydrogen discharge lamp. The weak 410 nm line is barely visible, but it shows that this old instrument is capable of resolving in this range. The LED light source also is deficient in the blue range.

The Bunsen spectroscopy consists of 3 tubes: slit and collimator tube, viewing telescope tube, and scale projection tube. This specific instrument uses a prism as opposed to a diffraction grating for dispersion and thus, wavelength is not linear with position. However, it can be calibrated using gas discharge tubes. The projected linear scale then simply assists in applying the calibration. Given the known spectral lines, wavelength is related to the ln of the linear position scale values, as shown in the plot below. We performed a similar study several years ago to help a homeowner’s association assess lighting performance of street lights.

Signal analysis

If you routinely are performing measurements of noise in instrumentation or other high frequency measurements it makes sense to use modern computer based digital oscilloscopes and waveform analyzers. However, if these measurements are not routine, many times older analogue equipment combined with a digital camera and image analysis software such as Image-J (free) or Photoshop can be used to extract the data you need. Below is an example of the analysis of the ultrasonic distortion from Direct Stream Digital (DSD) recording of audio. The figure at the left shows a stack of mid 20th century electronic instrumentation. DSD was not even dreamt of when these instruments were designed. The right hand images are screen photographs of a Hewlett-Packard 120B oscilloscope display. The right top figure is the crest of a sine wave with a frequency of 1 KHz at 0 dBVU and 0.3 % harmonic distortion, as measured by a Hewlett-Packard 130B distortion analyzer. The lower right figure is after DSD128 digitization and conversion back to analogue. Note the high frequency noise at about 10 mv peak-to-peak carried by the original sine wave. This is an inherent issue with DSD and recording devices employ filters and noise shaping to make the problem inaudible. Digital image analysis shows the lowest frequency ultrasonic noise to be around 45 KHz, which is consistent with expectations for the DSD128 format. The total harmonic distortion in this signal is about 2 %. However, it is not audible.