For the second 6.5m Prefire and later firings we have installed pairs of LVDTs at 4 azimuths on the oven to measure the aluminum and hearth radii. We have also installed 2 pairs of LVDTs on the Inconel bands. In September 98 for the third 6.5m casting, we installed two more pairs of LVDTs on the Inconel bands. Viewing the results of these measurements is tricky for at least three reasons:
1) We use the rotation A/D board in computer v0 to record the data each minute. The oven menu software thinks these are rotation A/D voltages. So to read the current values, look in the menu for Zone R. They also don't display quite full precision on the menus.
2) The LVDTs measure voltages with 12 bit resolution from -10 VDC (0) to +10VDC (4096). For the LVDTs with 2-inch travel, this corresponds to 1 volt = 0.1 inches. Thus, the raw A/D readings must be multiplied by some scale factors. We normally want to convert this reading to radius of the hearth or to equivalent temperature of the aluminum plate or to length of the Inconel bands. The hardwired scaling factors for the hearth and aluminum plate LVDTs are set to give readings in degreesC as an equivalent aluminum plate temperature. This assumes a plate radius of 198 inches and an expansion coefficient of 24.5e-6 per degree C (MIC 6 tooling plate). The hardwired scaling factors for the Inconel band LVDTs give readings in degreesC as an equivalent band temperature. This assumes a pulley ratio of 0.3831, an angle of ~13 degrees from the band (for the 6.5m bands), a hot band length of ~492 inches, and an Inconel expansion coefficient of 11.5e-6 per degreeC. The scaling factors from ADU to the readings on the screen are listed in mirror$src/oven/oven.h They are 412 and 1618 degrees full scale.
3) The LVDT readings have an arbitrary zeropoint offset which depends on how they were mounted at room temperature. This cannot be removed by changing the normal ograph multiplication factors. To get rid of this offset you must set the appropriate LVDT offset values in the Rotation Menu BEFORE the firing begins (before the data is collected). If the zeropoint is screwed up, you may need to resort to the imarith tricks described in the older memo below. Especially the band LVDTs need to be mechanically set near the end of their +-10VDC travel before setting the software zeropoints.
To plot LVDT temperatures, use ograph and set info?="rClvdtN", where C is the computer number and N is the LVDT number. So the first LVDT would be "r0lvdt0" and the last would be "r0lvdt15". Premade scripts that work like ograph are available in the scopes.furnace package for plotting band and aluminum LVDT readings. (see below) ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Here's a way to get all the LVDT readings with full precision for a particular minute. Only engineers would ever want to do this.
wtext rotv920219[17:32,1439] STDOUT head-
Obviously the number on the rotv image changes according to the date. The 1439 number is the number of the minute in the day (1-1440). I.E. 1439 is 2 minutes before midnight. STDOUT can be changed to the name of another file if you don't want to print to the screen.
------------------------------------------------------------------------------ Now to subtract the zeropoint offsets, you must use imarith on the rotv..... data images. Do not do this as "pilot" because there is a large risk that you will destroy the data altogether.
From your personal directory (not log'd in as pilot):
imarith /home/pilot/rotv920401 - /home/hill/rotvzero rotv920401
will make a zeropoint corrected rotv..... image. You can now ograph this in the normal fashion.
------------------------------------------------------------------------------- Scripts from scopes.furnace$, glvdt.cl and gband.cl will overplot the LVDTs on the Z6, or Z3 temperatures. To use them say "epar glvdt" or "epar gband" after loading the scopes.furnace package.
Sample commands as used in these scripts look like:
ograph info1=ztmp zone1=z6 info2=etmp zone2=z6 \ info3=ztmp zone3=z4 info4="" append=no
# LVDT readings scaled to aluminum temperatures.
# Hearth and Aluminum radius at 350 degrees ograph info1="" info2="r0lvdt0" info3="r0lvdt1" info4="" # append=yes
# Hearth and Aluminum radius at 160 degrees ograph info1="" info2="r0lvdt2" info3="r0lvdt3" info4="" # append=yes
# Hearth and Aluminum radius at 340 degrees ograph info1="" info2="r0lvdt4" info3="r0lvdt5" info4="" # append=yes
# Hearth and Aluminum radius at 170 degrees ograph info1="" info2="r0lvdt6" info3="r0lvdt7" info4="" # append=yes
------------------------------------------------------------------------------- # aluminum hearth plate LVDT scaling delta_T = delta_LVDT / radius_plate / alpha_plate = delta_LVDT (in) * 206 (K/in) / 10 (V/in) = 20.6 (K/V) / 2048 (ADU/in) = 0.1007 (K/ADU) full scale = 412 degrees
# Location on Furnace for Hearth/Aluminum LVDTs th=350 #0 is hearth to Box 0-1, #1 is aluminum to Box 0-2 th=160 #2 is hearth to Box 150-1, #3 is aluminum to Box 150-2 th=340 #4 is hearth to Box 270-1, #5 is aluminum to Box 270-2 th=170 #6 is hearth to Box 0-3, #7 is aluminum to Box 0-4 ------------------------------------------------------------------------------- # Geometry of Inconel Band LVDTs
# Location on Furnace (both 6.5m and 8.4m) #8 #9 #10 #11 # computer th=130 th=160 th=320 th=340 # oven angle (1-7) (2-7) (3-7) (4-7) # wire label Box 150-3 Box 150-4 Box 270-3 Box 270-4 # driver box
# Measured 6.5m geometry (Hill & Lutz) 06AUG93 # cable angle from band at room temp, add ~2 degrees for 1180C 12.9 deg 11.8 deg 11.8 deg 12.1 deg +- 0.3 deg # distance from pulley tangent to band at room temp 31.75 in 33.25 in 34.13 in 33.5 in +- 0.25 in
# average angle at 500 C ~= 13 degrees, cos(13) = 0.974 # computer will assume they (angles) are all the same for now.
# Measured 8.4m geometry 26MAR96 # cable angle from band at room temp, add ~2 degrees for 1180C (nominally set for 0 degrees = parallel to bands) (8.4m bandseat was run with the 6.5m scale factors. MAR96)
# distance from pulley tangent to band at room temp (nominally 33 inches from the end of the band)
# pulley ratio = 1.863 / 4.863 = 0.3831 # hot length of band = total length - cold length = 510 in - 18 in ~= 492 in # Inconel expansion coefficient 11.5e-6 per degreeC
# correct for pulley ratio, cable angle and pulley wrap up. # pulley wrap-up ~= pulley_radius * delta_theta
delta_Band = delta_LVDT / 0.3831 / (0.974 - 0.0178) = delta_LVDT * 2.73
# Did not correct for expansion of tub or expansion of furnace frame. # Did not correct for the inflection points in the Inconel expansion.
delta_T = delta_Band / Length_Band * 2 / alpha_band delta_T = delta_LVDT (in) * 965 (K/in) / 10 (V/in) = 96.5 (K/V) / 2048 (ADU/in) = 0.471 (K/ADU)
# full scale = 1930 degrees (probably good to +- 5%)
# 08SEP98 New Inconel Band LVDTs # Location on Furnace (both 6.5m and 8.4m) #12 #13 #14 #15 # computer th=50 th=70 th=260 th=280 # oven angle Box 60-1 Box 60-2 Box 60-3 Box 60-4 # driver box
Both the old and the new Inconel Band LVDTs are within a few degrees of being parallel to the bands. Nonetheless, we are still running with the original 6.5m scale factors. LVDTs 12-15 are scaled to be 50.8 (mm) full scale. Multiply by 38 to get nominal degrees K. (In the future they should be scaled to 1930 to match the other band LVDTs.)
Be sure to set all the band LVDTs to a position corresponding to -9.8 VDC to assure that the have plenty of travel to accomodate band expansion.
---