In this section we only briefly point out that the MARX-ACIS model cannot reproduce the degree of pileup in XRCF tests. In Figure 21 we show the PH spectrum for XRCF H-IAI-CR-1.001 (solid line) plotted against three realizations of the SAOSAC+MARX model for the nominal XRCF flux, the flux required to match the counts in the Al-K feature, and the flux required to match the counts in the first pileup peak. This figure and Table 6 show that no incident flux combination can reproduce simultaneously the main peak and pileup peaks. This problem had been encountered in an ongoing effort by McNamara & Wise to develop an algorithm to model pileup in Chandra science observations. However, prior to this point we were unable to determine conclusively the relative contributions of the mirror model and ACIS model to the discrepancy. Now the mirror model looks encouraging, while the detector model appears suspect.
| Data | BND Flux | Pileup | Peak |
| p cm-2 s-1 | (event) | (cts/chan) | |
| H-IAI-CR-1.001 | 0.00452 | 0.080 | 1499 |
| MARX | 0.00452 | 0.136 | 1297 |
| MARX, flux + 15% | 0.00520 | 0.157 | 1445 |
| MARX, flux-20% | 0.00362 | 0.112 | 1074 |
| MARX, flux-40% | 0.00271 | 0.084 | 834 |
