-
Notifications
You must be signed in to change notification settings - Fork 3
Problems
This page describes some problems related to S4ACS and the EMCCD cameras used in SPARC4.
On May 5, an issue was identified in which flat-field images acquired by the four SPARC4 cameras reached a saturation limit of approximately 5,000 ADU, whereas the expected limit should have been around 60,000 ADU. This behavior occurred without any error being reported by the devices, making the diagnosis considerably more challenging.
During the investigation, I determined that the problem was related to an incompatibility between the cameras and version 2.104.30158.0 of the SDK provided by the manufacturer. This library was used in version v1.56.1 of the S4ACS software, which was deployed during the nights of May 4 and 5. The first night was dedicated to engineering activities.
The issue was resolved by downgrading the SDK to version 2.104.30000.0, which had been used previously and did not exhibit this behavior.
After reviewing the manufacturer's release notes, I could not find any explicit indication that SDK versions newer than 2.104.30000.0 would be incompatible with iXon Ultra 888 cameras manufactured in 2016, suggesting that the Oxford Instrument company is unaware about this behavior.
The issue sporadically reported by the S4ACS systems regarding the effective exposure time of SPARC4 images was investigated. In this issue, the S4ACS sets an error flag in the image header whenever it detects that the difference between the effective exposure time of a newly acquired image and the configured exposure time exceeds 5%.
In this analysis, the images acquired on the night of 2026-06-21 were used, which was the most recent occasion on which the problem was identified. For each pair of images whose ACSERROR keyword was set to true, the time difference between their DATE-OBS timestamps was calculated, taking the exposure time into account. The result obtained was 4.5 ms, which corresponds to the image transfer time between the acquisition and readout registers of a frame-transfer device. This value is significantly different from the exposure-time discrepancy calculated by the S4ACS when it sets the error flag.
Based on these findings, the source of the problem is likely related to the timing information retrieved by LabVIEW (which is independent of the cameras' internal timing control) during the calculation of the effective exposure time. In contrast, the timing information recorded in the image headers appears to be unaffected by this issue.