Low-resolution

Single Raw Frame

Low-resolution Spectroscopy
						count rates (ADU/s/coadd)
Filter	source	sp. type	mag.	itime (s)	slit (")	source	sky line	bkg
NIRSPEC-2	HD 89239	A0V	V=6.528	4	42 x 0.76	3500	9	2
NIRSPEC-3 (J)	HD 129653	A2	J=6.980	3	42 x 0.76	3900	12	-
NIRSPEC-5 (H)	FS 27	G8IV/V	H=13.181	5	42 x 0.76	13	25	1
NIRSPEC-7	FS 27	G8IV/V	K=13.123	5	42 x 0.76	14	16	1-5

source and sky line count rates are peak
bkg count rates (between sky lines) are typical unless there was a large range (i.e. in NIRSPEC-7)

Background Subtracted Nod Pair (using positive beam only)

Low-resolution Spectroscopy
Filter	source	sp. type	mag.	itime (s) x coadd	slit (")	SNR/pixel
NIRSPEC-5 (H)	FS 27	G8IV/V	H=13.181	5 x 2	42 x 0.76	12
NIRSPEC-7 ~2.0µm	FS 27	G8IV/V	K=13.123	5 x 2	42 x 0.76	12
NIRSPEC-7 ~2.2µm	FS 27	G8IV/V	K=13.123	5 x 2	42 x 0.76	8
NIRSPEC-7 ~2.4µm	FS 27	G8IV/V	K=13.123	5 x 2	42 x 0.76	3

by using the negative beam, one can increase SNR by root 2 (1.4) in the source noise-limited case.
Decreasing source counts with increasing wavelength and increasing background noise with increasing wavelength conspire to drive down the SNR in NIRSPEC-7 at longer wavelengths.

High-resolution

Single Raw Frame

High-resolution Spectroscopy
							count rates (ADU/s/coadd)
Filter	source	sp. type	mag.	itime (s) x coadd	slit (")	seeing (")	source	sky line	bkg
NIRSPEC-7	HR 3067	A3V	K=4.362	4 x 1	0.576 x 12	˜0.5	1760	n/a	0
NIRSPEC-7	HR 1237	A0V	K=6.074	15 x 1	0.576 x 12	˜0.5	410	n/a	0
NIRSPEC-7	HIP 53735	A0V	K=8.856	150 x 1	0.576 x 12	˜0.5	32	n/a	0

seeing estimated from FWHM across spectrum of 3.5 pixels (x 0.144 "/pixel)
source count rates are peak
the few sky lines were not measured
average backgrounds were near 0

Background Subtracted Nod Pair (using positive beam only)

High-resolution Spectroscopy
Filter	source	sp. type	mag.	itime x coadd x frames	slit (")	typical counts	SNR/pixel
NIRSPEC-7	HR 3027	A3V	K=4.362	4 x 1 x 2	0.576 x 12	6000	150
NIRSPEC-7	HR 1237	A0V	K=6.074	15 x 1 x 2	0.576 x 12	6000	150
NIRSPEC-7	HIP 53735	A0V	K=8.856	150 x 1 x 2	0.576 x 12	4000	120

by using the negative beam, one can increase SNR by root 2 (1.4) in the source noise-limited case.
typical counts is "St" in the below relation.
rms in background (B) was 5-15 in all cases (assumed 10 ADU)

SNR/pixel values were determined by the relation:


                 Sgt                    S = source count rate (ADU/s)
SNR =   ______________________   where  B = rms in background (ADU); implicit time dependence
        ______________________          D = dark current (0.2 e-/s/pix)
       /                                g = gain (4 e-/ADU for SPEC)
      / Sgt + Bg + Dt + r^2             t = integration time (itime x coadds)
    \/                                  r = read noise ( 2 x (25 e-)^2 = 1250 e- for nod pair)

The calculated values match a "by-eye" assessment using cuts through the spectra.