Notes on planning high precision relative photometry observations Planning observations of transiting exoplanets, studies of brown dwarf weather, and other investigations requiring high precision relative photometry (HPRP) with IRAC is different than most IRAC observations described in Spitzer Observing Manual. For the HPRP observations, it is important to maximize source signal-to-noise while minimizing systematic effects; therefore, observers should integrate with the longest possible frametime that does not saturate (stellar photometry mode is very useful in these cases) and collect data in staring mode (no dithers) with as many repeats as needed to cover the desired duration (e.g. transit, rotation period). For bright sources (2-3 times the published saturation limits) which may saturate in even the shortest provided integration time, the observations can still be done, but the source will need to be centered at the corner of a pixel to spread the flux and place the center pixels below saturation. To place a source at the corner of a pixel use a fixed cluster target with array offsets of 0.6 arcseconds in each direction and select the "observe offsets only" option. As Spitzer exhibits a pointing wobble of amplitude ~0.15 arcseconds and period ~3000 seconds, the measured flux densities at 3.6 and 4.5 microns will vary due to the pixel-phase effect. While on average, the pixel-phase effect can be corrected for as described in the IRAC Data Handbook, each array pixel has a slightly different effect and observers should trend and remove the effect from their own data. As a result, observers should plan to spend equal times on and off eclipse in their observation and have at least 3000 seconds of baseline to cover a wobble period. In addition to the wobble and subsequent variations in photometry due to pixel phase, staring observations have shown that IRAC has relaxation effects that cause the observed photometry to vary with time by several (and up to 10) percent. At 8.0 microns, the flux density increases nonlinearly with time as shown in Knutson et al. (2007). At 3.6 and 5.8 micron, there appears to be a decrease with flux density (Beaulieu, private communication). The arrays mostly stabilize within an hour or so. It is believed that the relaxation effects depend on the incident source flux density as well as the history of what those pixels have observed. There is some evidence that observing a bright source, "flashing the array", before the HPRP observation mitigates this effect (Deming, private communication). A dedicated test of this mitigation technique will be run in an upcoming IRAC campaign; however, we have no recommendation at this time on the efficacy of array flashing. We suggest that observers pad the start of their observation by an hour or so to minimize the relaxation effects. For observations longer than eight hours, an instrument engineering request (IER) will have to be generated for accepted observations. In preparing the proposal, please mention that you will require a staring observation of more than eight hours which requires an IER. Please create AORs that span the desired duration and use a follow-on constraint to group them. Provide your desired visibility windows in all cases. Due to high data volumes of some of these observations, your observations may be converted to IERs which remove data taking in the off-source fields of view. The SSC will consult with the PI's and technical contacts of approved programs when this is required.