Stars & the cosmos · START WITH SOMETHING FAMILIAR
A blackbody made for comparison
Why does a sky measurement need a laboratory reference?
FIRAS’s external calibrator was a shaped absorber made from iron-loaded epoxy. It could fill the instrument’s sky aperture, giving a controlled reference observed through the same measurement system. Its reflectance, temperature gradients and leakage had to be characterized. The iron and carbon links describe the reference material; they are not chemical species inferred from the cosmic spectrum.
The original analysis includes frequency correlations. The published approximation gives neighboring bins a correlation of 0.176 and bins two steps apart −0.203.
These are published observations after calibration and foreground modeling. NASA’s intensity column combines a 2.725 K reference blackbody with measured residuals. Moving temperature changes only the comparison model.
Published bins ± pointwise 1σCalculated Planck curve
The intensity scale is fixed. The small uncertainties become visible in the residual plot below.
The differences reveal more than the peak
Published bins ± pointwise 1σCurrent reference-relative comparison
The vertical range adapts to the current and saved comparisons. Error bars describe the archived bins, which are correlated.
Descriptive unweighted RMS0.07154 MJy/sr
Published bins43
Archive reference2.725 K
RMS summarizes the displayed reference-relative differences. It is not a fit probability or a temperature confidence interval: neighboring bins are correlated, and the archive already uses a fitted reference.
Archived intensity
383.478 MJy/sr
Archived residual
-10 kJy/sr = -0.01 MJy/sr
Pointwise uncertainty
±18 kJy/sr (1σ)
Current reference-relative difference
-0.010000 MJy/sr
How the comparison is constructed
r(T) = r_archive + Bν(2.725 K) − Bν(T)
Residuals and errors in the source table are kJy/sr; divide by 1000 for MJy/sr. We preserve the archived residual exactly at 2.725 K. Recalculating the original baseline with modern SI constants at the printed wavenumbers differs by up to 0.007598 MJy/sr, so this is explicitly a reference-relative comparison.
Bν = (2hν³/c²) / [exp(hν/kT) − 1], with ν = 100cw for w in cm⁻¹. The galaxy template is not subtracted again. The later published 2.72548 K temperature combines other evidence and is not a value this table independently recovers.
Read all 43 original rows and comparison values
cm⁻¹
Intensity MJy/sr
Residual kJy/sr
1σ kJy/sr
Galaxy model kJy/sr
Current difference MJy/sr
2.27
200.723
5
14
4
0.005000
2.72
249.508
9
19
3
0.009000
3.18
293.024
15
25
-1
0.015000
3.63
327.77
4
23
-1
0.004000
4.08
354.081
19
22
3
0.019000
4.54
372.079
-30
21
6
-0.030000
4.99
381.493
-30
18
8
-0.030000
5.45
383.478
-10
18
8
-0.010000
5.9
378.901
32
16
10
0.032000
6.35
368.833
4
14
10
0.004000
6.81
354.063
-2
13
12
-0.002000
7.26
336.278
13
12
20
0.013000
7.71
316.076
-22
11
25
-0.022000
8.17
293.924
8
10
30
0.008000
8.62
271.432
8
11
36
0.008000
9.08
248.239
-21
12
41
-0.021000
9.53
225.94
9
14
46
0.009000
9.98
204.327
12
16
57
0.012000
10.44
183.262
11
18
65
0.011000
10.89
163.83
-29
22
73
-0.029000
11.34
145.75
-46
22
93
-0.046000
11.8
128.835
58
23
98
0.058000
12.25
113.568
6
23
105
0.006000
12.71
99.451
-6
23
121
-0.006000
13.16
87.036
6
22
135
0.006000
13.61
75.876
-17
21
147
-0.017000
14.07
65.766
6
20
160
0.006000
14.52
57.008
26
19
178
0.026000
14.97
49.223
-12
19
199
-0.012000
15.43
42.267
-19
19
221
-0.019000
15.88
36.352
8
21
227
0.008000
16.34
31.062
7
23
250
0.007000
16.79
26.58
14
26
275
0.014000
17.24
22.644
-33
28
295
-0.033000
17.7
19.255
6
30
312
0.006000
18.15
16.391
26
32
336
0.026000
18.61
13.811
-26
33
363
-0.026000
19.06
11.716
-6
35
405
-0.006000
19.51
9.921
8
41
421
0.008000
19.97
8.364
26
55
435
0.026000
20.42
7.087
57
88
477
0.057000
20.87
5.801
-116
155
519
-0.116000
21.33
4.523
-432
282
573
-0.432000
Data provenance
Every non-comment row from NASA’s released table, in original order. No points were added or removed. The download page dates delivery to March 2003; the file header dates its initial revision to May 2005. Both are preserved in our source record.