HUG Water Flow Calorimeter V1.0

for LENR-Stick Test Cells

Prepared by:  

Ryan Hunt, R&D Manager

Paul Hunt, Chief Engineer

This newest calorimeter was made to improve upon the Concentric Tube Calorimeter but still accommodate the same test cell within it.  Our goal was to make something that was simple to understand, at least as accurate, easier to build, and overall, more copyable.  This document is meant to introduce the novel device we have come up with.

The basic idea is that the LENR-Stick Test Cell is inserted into the middle of a copper tube that makes up the water jacket.  There is an air gap that allows the LENR stick to operate at high temperatures (500+ C) while the water jacket remains at a comfortable 30C to 60C range.

The water feeding the water jacket comes from a temperature controlled reservoir with a small, submersible pump in it.  The water arrives at the calorimeter within one degree of the set point, and usually much closer.  The water is then precision controlled and metered and used to cool a copper tube around the device under test.  Instead of measuring the exact mass flow rate, or the volume flow rate, however, we measure the heat capacity flow rate.

This whole design represents a new tool set that HUG can now rapidly adapt to other calorimetry tasks and experimental arrangements.

Core of Apparatus:

Image of whole apparatus:  Below, we see the foam enclosure, the water reservoir beneath it, the solid state heat pump next to the reservoir, and the PC to administrate it.  

Under the first layer and within the bottom layer is an assembly of aluminum plates and water heat transfer plates that help isolate the internal parts from any ambient temperature changes.

The actual calorimeter core is within this block of foam and is accessed by lifting the layers off.

Copper Calorimeter Body:  Made from a heavy wall ¾” OD copper pipe, and a ¼” OD copper tube soldered to it.  The thermal conductivity of the thick copper tube makes sure that the heat flows rapidly to the water in the tube.  The small volume of water in the tube helps keep the time constant relatively short.  The flow rate is roughly 15 mL/min.

Graph of steady state thermistor readings with no power to device under test.  

• Therm_1 (red line) shows the incoming water temp from the reservoir and that it varies by 100 mK, or so.  
• Therm_2 is the stabilized input temp at the target of 20.5C.  It varies by roughly +/- 2 mK in the short term average.  The longer term averages show +/- 1mK
• Therm_3, after the  metering heater, shows it exactly 1K higher at 21.5C +/- 2 mK
• Therm_4, after the copper shell appears lower, but a good clean signal.  It is lower because of heat loss through the foam insulation to the aluminum ambient isolation plates above and below.  The aluminum plates are stabilized with water at the same temperature as Therm_1, or roughly 1 to 2 degrees lower than the water passing through the calorimeter.

Solid state heat pump assembly with the capability to handle cooling for up to 200W of experimental load.  This one reservoir and heat pump may be suitable for supplying controlled water for several experiments simultaneously

The reservoir is made from a 5 gallon plastic pail with a screw on lid to prevent evaporation.  In the image below, the hoses to the left go to the calorimeter and the hoses on the right go to and from the heat pump.  Each pump flows roughly 8 l/min at between 1 and 2 meters of head pressure.

On the inside of the reservoir, the pumps sit on the inside of a ring of filter material.  The return water is delivered on the outside of the filter material.

Details of thermistor and heater assemblies

Tubing is ¼” OD silicone tubing.  The fittings are plastic drip irrigation fittings.

Heater is a series of (8) ⅛ watt, 18 ohm encased in Kynar heat shrink tubing and hot melt glue.  

In the picture

The heater is sealed into the fittings with hot melt glue.

Temp sensors are after right angles so we get extra stirring from the flow around a square corner.

The thermistor is enclosed in Kynar heat shrink and hot glue, and directly in the water flow.  They are 10K Ohms@25C, 0.5% tolerance.  (model cantherm MF51E103E3950  Digikey.com part# 317-1305-ND)

UPDATE 8/14/13 -- New thermowell design implemented

Below:  Tiny copper thermowells made from .052" ID copper tube crimped and soldered on one end.  They insert through the wall of the silicone tubing.  They seal by themselves, but we are adding a fine film of silicone seal to help make sure the seal is tight.

Below:  A new thermowell inserted for the Thermistor 3 sensor.

Below:  We see Thermistor 4's thermowell going into place.  Beside it is the exposed trouble maker of the day wrapped in the heat shrink tubing.  I took the sensor under a microscope, but didn't see any obvious signs of corrosion or any visible moisture with the plastic tubing.  

Below:  We see Thermistor 2 being installed with a dot of silicone and the cotton thread that will act as the wick.

Unfortunately, this means the test will be delayed till next week as the silicone adhesive cures over the weekend.  We are hopeful that as long as both manners of installing the sensors provide accurate reading of the actual water temperature that it will not affect the calibration.  That will have to be tested, though.  We also, do not believe that the sensors were drifting during the calibration runs because of the tightness of the many different runs we did through the calibration steps.

It turns out that the coefficients on the thermistors varies slightly and it will require recalibration.

END UPDATE --------------------------

Update as of 08/16/2013

We have changed our over temperature sensor #3 (on the copper tube).  We had  trouble with the alarm shutting everything off at 26 watts.  We took a piece of Al flashing and put 12 dimples in it.  Then we bent it around the copper tube and taped the sensor to the Al .

This should eliminate any local hotspots for the sensor.

End Update-----------------------------

LENR-Stick for Multi-Wire test:

Calibration Procedure

While we get a pretty good idea of the power out just by the rough correlation of a water flow rate of 1 watt/degree of temperature rise.  However, there are parts tolerances for the thermistors and the current shunts that may make that inaccurate, and some heat is lost through the insulation. Many of these possible sources of error are constant and calibrate out, nicely, though.

First, zero calibration

-Get all 4 thermistors at the same temperature (turn off all the heaters and just run the water and let it all come to equilibrium.  Then we run  a script that takes the latest 5 minute average and adds offsets to make the thermistor readings the same.  This is not essential, but it is convenient to see accurate relative temperatures.

System Calibration

1. Vary the input power in steps
2. Record the final temperatures and indicated rough power out
3. Do a curve fit of the resulting data
4. Add a calibrated output to the data columns using the curve fit equation.

(In the future, the entire calibration process may be scriptable)

Results of Calibration

From six total calibration runs we have the results in table form below.  The 95% CI ranges from 20 mW to 200 mW depending on the input power range.  The 99% CI ranges from 30 mW to 310 mW.  

 Full spreadsheet here

We will translate the DUT output to the calibrated output power based on the below graph.