having promised some data, . . . .
here is some describing the search for reality in calibrating a new test bench with commercial equipment, as contrasted with Incoherent rolling his own
For temps I started with some Keithley and Fluke RTD thermometers with 0.01C resolution, to which I added some ¼” 1/3 DIN B tol 4-wire sensors. Sent it off to the cal shop and got them back with nice certs, which did not seem to jive too well with my comparisons. So I accepted the thermometers as being ‘in spec’ per the certs, but would have to recal for more accuracy using a Haake A82 recirculating chiller with a HP 2804A quartz thermometer with 2 probes. The bath was ‘mapped’ and the probes then caled and cross checked, with the results shown in the RTD cal graph.
(Thanks for the help Les)
essential to this activity is the mapping of the bath at higher resolution, and the consistency
The next task was to quantify the head loss of the instrumentation fittings and the connecting tubing to the DUT. The crosses are (now) as shown in the photo, but were connected together for testing with a single piece of tubing. Using the 823DP data, a head loss curve was developed, but a temperature anomaly was noted; there was an ‘apparent’ temperature drop across 6” of tubing with no temperature difference between the coolant and ambient.
The connection temperature offset graph shows several crosses tested several ways, and my first inclination (and action, alas) was to use the Series 5 data as an offset of 0.054°C; more reflection suggests that the Series 4 equation is a better description.
temp related stuff is delayed ‘till I redo the spreadsheets, etc.
the cross head loss is as shown, but will be redone at a higher flow resolution of 0.001gpm.
so at the moment all I have is a head loss curve for the CoolingWorks CoolRad-12T
more later
BTW, "C" will be coolant inlet less air inlet
|