As far as I can see in tempermeters we use:
1. Aluminium cup is filled with sample chocolate, this is placed in "cup holder", covered, temperature probe is inserted in the centre of the sample and test is started.
2. Tempermeter provides consistent cooling (by controlling the temperature of "cup holder" at +8C).
3. Temperature recorded by the probe is recorded as it the sample is cooled and solidified.
4. This data is plotted and analysed.
5. If chocolate is perfectly tempered and monocrystaline structure is formend during crystalisation time temperature curve will be: down, stay at constant temperature and down again (close to how it would look for freezeing water) and temper index will be calculated as 5.
6. Undertempered and overtempered chocolates produce different time temperature curves (there is no "stay at constant temperature" part of the curve).
7. We agree on acceptable temper index range and if sample tests show that chocolate is within the range we start production, keep testing through the day and adjust tempering settings (on tempering machine) if needed.
Making one yourself could be done. As the whole test takes 8 minutes no more than 3-5 recordings per second would be needed (maybe even less). Containers are aluminium coffe capsule cups. Consistent "cup holder" temperature can be achieved by flowing constant temperature water through the coil), large tank and basic cooling control system will achieve this. Probes, hardware and software can be easily done as Kevlarcoated above stated.
For somebody with knowledge and time this is not so complex.