Rossi Lugano/early demo's revisited. (technical)

  • Length of heater wire in the rods


    A few weeks ago I did some analysis using data from the MFMP thermal dogbone test based upon the adapted emissivities which I found to limit the errors between thermal and applied powers.

    I assumed two different length of heater wire in the rods, 4 cm and 6 cm.

    These leads to two different heating distributions and thus also to two different power densities under the ribs.

    Based on the found power densities I recalculated from the MFMP data the temperatures of zones 5 through 9 if the emsissivities would have been those of alumina.

    For 4 cm of heating wire in the rods I calculated a temperature of 510.69 degree C.

    For 6 cm of heating wire in the rods the temperature found was 440.78 degree C.


    Since the temperature of 440.78 for the 6 cm heater wire in the rods is the closest to the reported 449.9 degree C in the Lugano report, it is an indication (as far as my assumptions about the MFMP test hold) that indeed the length of the heating wire in the rods is probably more then the 4 cm assumed.


    Note also that the if we assume inflated temperatures due to using broadband emissivities on the Optris, that the average temperaure for zones 5 through 9 in Lugano would have been 376.7 degree C.

    This value is largely outside the temperature range of 441 through 511 found by recalculating the MFMP data for different heater wire lengths in the rods.

  • Please could you make two more spreadsheets in the same form factor as the Dummy one, (I realize that this quite onerous), for both the recalculated lower temperature Run 16 and your COP 5 version of Run 16 ? Then it will easier to see what the simulation is doing. The Ribs can be treated as a single section for now. I am more interested in chasing why ~ 24% of the electrical input goes into the Rods and Caps, but ~ 50% goes out, when ~76 % of the input power goes into the thin Ribs section in the Dummy cases.


    I completed as per your request the spreadsheet for the COP 5 version.

    See attachment.


    I still need to make a start on the one for the recalculated lower temperatures.

    (Working now on how to get the temperatures of the rods, most likely a separate post)

  • LDM ,

    I have been going over the construction of the MFMP Dogbone replica of the Lugano device.

    An important aspect is the lead passing through the Cap. The '3 phase' version (3 windings, but wired in series outside of the caps due to single phase power supply) failed before the arrival of the Optris, so only the single phase version was imaged by the Optris IR camera.

    Anyway, note that the Caps are essentially entirely heated by conduction from the Ribs area, since a copper sleeve was placed over top of each of the leads, crimped to the supply winding lead at the Ribs-Cap interface, and the supply cable clamped to the lead while inside that copper sleeve.

    (This was to prevent the lead from burning off, as was experienced with one of the leads in the three winding version)




  • I have been going over the construction of the MFMP Dogbone replica of the Lugano device.

    An important aspect is the lead passing through the Cap. The '3 phase' version (3 windings, but wired in series outside of the caps due to single phase power supply) failed before the arrival of the Optris, so only the single phase version was imaged by the Optris IR camera.


    Any reason why the Kantal wire failed ?

    In general Kantal (if produced by Kantal) is very reliable.

    Maybe Alan S has some suggestions ?


    Anyway, note that the Caps are essentially entirely heated by conduction from the Ribs area, since a copper sleeve was placed over top of each of the leads, crimped to the supply winding lead at the Ribs-Cap interface, and the supply cable clamped to the lead while inside that copper sleeve.

    (This was to prevent the lead from burning off, as was experienced with one of the leads in the three winding version)



    Since, as you also noted in one of your posts, the conduction, even for a reasable conductive material as Alumina, is quite limited sideways, the center temperature of the ribbed area is almost not dependent what is happening in the caps and to the caps rod interface.

    I see that also in my FEM simulations.

    So if you apply to the ribbed area the power density you want to test with, then the center rib area temperature can in my opinion still be used for comparative measurements.










  • LDM ,

    What I was hoping to assess was the temperature of the Caps relative to the Ribs with power applied within the Caps.

    Unfortunately there is no data for the Caps temperature until the single phase version was run.


    The Fat Coil version may have had more input to the Caps, but I haven't found any data for that version at all (yet).

    Is there any Optris data for the Fat Coil Glowiness test?

  • Unheated Caps results in large thermal losses from the outer Ribs area (especially areas 1 and 10, but in turn the adjacent segments all lose heat towards the Caps), resulting in an uneven heat distribution along the Ribs area. The Caps become large heat sinks.

    This results in the MFMP thermal state differences from the Lugano device in their report.


    The heat distribution in the Lugano Ribs area is remarkably flat. This is because the Caps are heated, as well as the Rod ends, which limits thermal losses outwards from the Ribs area.

  • Any reason why the Kantal wire failed ?

    In general Kantal (if produced by Kantal) is very reliable.

    Maybe Alan S has some suggestions ?


    Possible that the wire had been 'notched' a little during the winding process. Generally it is necessary to grip the end few feet of a wire very firmly while the last turns are laid down Excessive force or the wrong pair of grips (serrated jaws rather than smooth for example) can easily do this -and it can be overlooked. The damaged section experiences a higher current flow per unit of area (smaller cross section) and will run hotter than the undamaged sections and this can (and often does) cause premature failure.


  • What I learned from interpreting some simuations is that the following might be the case


    1. To get to the reported cap temperatures, the heater coil might have been running for some length also under the end caps


    2. You can not from radiation/convection get much power into the rods, thus almost all power to the rods needs to be delivered by the heater wire. 4 cm into the rods is in my opinion not enough.


    3. If most of the power to the rods is delivered by the heater wire, then the rods cause a kind of thermal barier near the end cap. This barier limits to a great extend the loss of power at the side of the end cap, thereby reasing the temperature in the end cap


    The question now is which heater configurations are fullfilling these requirements and have also the required total resistance.

    If it is limited to a few cases, then you can for each case calculate the power under the ribs.

    Then test your model with the lowest and highest power setting required for getting the same power under the ribs.

    It will give you the maximum and minimum center temperature of the ribs.

    You can then compare those with those in the Lugano report and those of the MFMP test.

    Maybe you can draw some conclusions from it.


  • The heat distribution in the Lugano Ribs area is remarkably flat. This is because the Caps are heated, as well as the Rod ends, which limits thermal losses outwards from the Ribs area.


    In my opinion there is another effect in play

    The ribs closest to the caps and the inner side of the heated caps are viewing each other for a significant part.

    Since both are at elevated temperature, the heat exchange of the involved areas becomes much smaller.

    This smaller heat exchange must be compensated by a higher temperature in order to get the power out.

    This will result in that the temperature of the ribs close to the inner side of the end caps will be raised.

    (for the ribs farther away from the end caps the effect is also present, however to a less extend)

    This might also an explanation for the flatter profile.


    That the MFMP test had a more parabolic profile is also due to a much larger heater wire length under the ribs compared to the heater wire length under the caps then was the case for the Lugano ECAT.



  • I didn't get into the exact process, but the directly heated Caps are ultimately reducing the heat flow from the Ribs towards the Caps, for several reasons, including conduction, radiation, and advection.


    That the MFMP test had a more parabolic profile is also due to a much larger heater wire length under the ribs compared to the heater wire length under the caps then was the case for the Lugano ECAT.


    The length of the wire is not so important (as long as there are no huge gaps between wraps) as much as the heat power made by the length of the wire. The wires in the Lugano device were twisted pairs of 15 Ga Kanthal, which has a cross sectional area of about 3.3 mm2 (0.55 ohm/m), compared to the MFMP version with 18 Ga wires, with a cross sectional area of about 0.82 mm2 (1.76 ohm/m) . This is why in Lugano the hugely expensive power controller was required. The current flow in Lugano was very high (presumably in very short pulses) compared to the MFMP devices, for the equivalent Joule heat output per meter of wire. At around 450 C external, central Rib temperature, the MFMP RMS current for their device was around 6 amps, while the Lugano device required almost 20 amps to achieve the same central external Ribs temperature, due to the very low resistance. (The resistance of 1 m of Lugano device twisted wire is similar to that of about 10 m of coat hanger wire.)


    Other than the "three phase" version, and possibly the Fat Coil version, the Caps had effectively no direct heating at all in the MFMP versions.

    The resistance of the copper tube crimped onto the heater coil lead wire as it passes through the each Cap is so low that Joule heating is nearly zero there.


    In the case of the Lugano device, at least 12 cm of actively heated wire was passing through each Cap. (compared to effectively none in the MFMP versions).

  • Rod temperatures of Lugano active run period 16 if broadband insted of in-band emissivities on the Optris where used


    In order to be able to calculate the lower rod power of the Lugano active run period 16 in the case that the measured temperatues where inflated, we need to know the rod temperatures.


    However only the total convective and the total radiated power of the rods are given.

    To get an approximate temperatures we have three possibilties to determine average temperatures.


    - Iterate between h and the temperature till the reported convective power is reached

    - Iterate between the emissivity and the temperature untill the reported radiated power is reached

    - Iterate with both changing h and emssivity temperatures untill the total power is reached


    For Lugano active run 16 the following data was given for the power of one set of three rods


    Convective power--------87.94 Watt

    Radiated power-----------88.47 Watt


    However the Lugano testers applied afterwards a factor 2/3.

    Before continuing we need to undo this by multiplying both the convective and radiated heat by 3/2

    The corrected powers become


    Convective power--------131.91 Watt

    Radiated power-----------132.71 Watt

    Total power-----------------264.62 Watt


    With the above corrected data all three types of iterations where done with the following results


    --------------------------------------------Temp----convective power----radiated power----total power

    -----------------------------------------------C---------------Watt-------------------Watt-------------------Watt

    h iteration----------------------------120.28----------131.90------------------95.08----------------226.98

    emissivity iteration----------------145.33----------173.04----------------132.70----------------305.74

    h and emissivity iteration-------132.55----------151.87----------------112.75----------------264.62


    As expected none of the iterations get both the convective and radiated power about equal to the reported powers.

    This is due to the fact that the convective power is a linear function of the temperature, while the radiated power depends on the 4th power of the temperature (in Kelvin).


    For a recalculation it is proposed to use the combined h and emissivity iteration since the power is equal to the calculated power and the temperature calculated is about the average of both the h iteration and the emissivity iteration.

    It must however be understood that the found temperature is an approximation.

    Using the temperature of 132.55 degree C with its associated broad band emissivity of 0.696 the recalculated average rod temperature in case broadband instead of in-band emissivities where used on the Optris becomes 111.17 degree C ( n = 3.938).


    The found average temperature of 111.17 degree C for the rods can be used to calculate the approximate power of the rods in case the temperatures of the Lugano active run period would have been inflated by using broadband emissivities on the Optris.

  • The 2/3 Rod adjustment factor is not explicitly shown to be applied to the Active Runs in the report.

    The convective and radiant powers are shown in a short table after a brief discussion, and must be considered to be each to have been adjusted already by the 2/3 factor before being listed if the 2/3 was indeed applied. However, the report does not indicate if this was the case or not.


    For the Dummy run, the convective and radiant power were first totalled, then the 2/3 factor was applied.

  • Logically, the amount of heat required to bring the lower-temperature outer Rib segments up to the same temperature as the central higher-temperature Rib segments should equal the power dissipated in the Caps, when the Caps are otherwise unheated.

  • The 2/3 Rod adjustment factor is not explicitly shown to be applied to the Active Runs in the report.

    The convective and radiant powers are shown in a short table after a brief discussion, and must be considered to be each to have been adjusted already by the 2/3 factor before being listed if the 2/3 was indeed applied. However, the report does not indicate if this was the case or not.


    For the Dummy run, the convective and radiant power were first totalled, then the 2/3 factor was applied.


    It is indeed correct that this is not explicitly shown in the report and your further remarks are to the point.


    So i propose based on my previous post to base the calculation of Lugano run period 6 for the case the temperatures where inflated on the presumption that the factor 2/3 was applied.

    But in addition to that I will also calculate the average rod temperatures for the case the factor 2/3 was not appllied and based on that temperature also the rod powers.

    It will tell us if this makes a large difference for the calculation the active run period 6 if the temperature wold have been inflated.

  • My general plan for the Caps test unit is to power it with 3 sets of folded-in-half and parallel (not touching, “bobby pin”), 30 cm long 24 AWG wires. (15 cm long folded), which will be run in series. Folded and passing through the 4x4 cm Cap along a 0.5 cm radius of the cylinder axis, the wire connections will be at the end of a nub representing a section of Rib type. 5 cm of the folded end of the pairs will extend out the opposite side, to a Rods area. (I was thinking of making an attachment modification so that the previous Cylinder hot tube could be used to simulate the Ribs, but this makes things unnecessarily complex.) This will be a roughly 8 ohm resistance once assembled, so rather sensitive to the voltage, and will probably melt the wire if over 70 V AC is applied.

  • This is intended to work with Rods, but I haven’t made immediate plans for them yet.


    I am currently working out the best casting method to get all the wires in place and a tiny hole down the center for a moveable thermocouple. I probably need to get another box of thermocouples and mini connectors since I have so many thermocouples cast into other things. I scavenged the last bunch from yet earlier things, and the wires are getting quite short now.

  • Rod temperatures Lugano active run period 16 if factor 2/3 was not applied


    When the factor 2/3 was not applied to the rod power, we have the following data


    Convective power--------87.94 Watt

    Radiated power-----------88.47 Watt

    Total power---------------176.41 Watt


    Iterating with both the convective heat tranfer coefficient h and the emissivity on the total power of 176.41 watt for one set of three rods gives the following result


    --------------------------------------------Temp----convective power----radiated power----total power

    -----------------------------------------------C---------------Watt-------------------Watt-------------------Watt

    h and emissivity iteration-------103.18----------104.84-----------------71.57-----------------176.41


    Using the temperature of 103.18 degree C with its associated broad band emissivity of 0.684 the recalculated average rod temperature in case broadband instead of in-band emissivities where used on the Optris becomes 85.72 degree C ( n = 4.171).


    When making the spreadsheet for Lugano dummy run period 16 in case the temperatures where inflated also a rod power calculation for the temperature of 85.72 degree C will be included.

  • Lugano active run period 16 recalculation if temperatures where inflated.


    The attached spreadsheet contains a recalculation of the Lugano active run period 16 if the temperatures where inflated due to using broadband instead of in-band emissivities on the Optris thermal camera.

    (Calculation when temperatures where not inflated was published in post #522 )


    The recalculation assumes that a factor of 2/3 was already included in the reported rod powers.

    However since it was not expliciy stated that the Lugano testers applied this factor 2/3, the sub page for the rod powers in the spreadsheet also includes the calculation for the case that the factor 2/3 was not applied.

    Note also that since the calculations are based on average temperatures, the results must be interpreted as approximate values.


    The recalculation with the established lower temperatures results in a total convective and radiated power of 1329 Watt assuming the factor of 2/3 was applied to the reported rod powers.


    If the factor of 2/3 was not applied to the rod powers the total calculated power would have been 1258 Watt


    The total applied electrical power for this run was 865 Watt.